Power supply design (2)

[liquibyte]

The output should never rise above the set point when power is applied or removed.  I am inclined to believe that there is more than one problem but the lack of any regulation or protection on the supply to the operational amplifiers I consider a design flaw.

That's kind of what I was thinking.  The design has never really worked from what I can tell.  Reading the lengthy thread(s) on the design gives people that have had success building the thing but I don't think anyone has really tested it until very recently.  One of the members over there just recently got a DSO and tried to solve the transient issue with the addition of the voltage regulator but I think he only tested within a certain range and reported it solved.  It wasn't.  I'm about ready to give up on this one for awhile and see if I can get something better going with a bit more research.

Any word on a vetted design that would be relatively simple that I could perhaps use the two 28V 4.3A transformers with?  I was shooting for 30V @ 3A but I also just got, in addition to the 301A's, OP07's and TL071's, a single primary, dual secondary transformer that can do 36V @ 1A series, 18V @ 2A parallel or 2X18V @ 1A independently.  I was thinking 0-18V @ 1A positive and negative until I can figure out this other issue, I just need a proven design that won't set me back too much.

The PS-503 design is pretty dense and unconventional in some respects even if you ignore the complexity added by being a dual output design with tracking.  There are changes I would make like having the output amplifiers operate with a fixed gain and varying the output from the reference because it would make the frequency response more predictable.

The PS-503 operational amplifiers are just 36 volt 741 types and not critical.  Besides qualifying their sources, Tektronix graded incoming 741s and the ones with high bias current, high offset current, high offset voltage, and high offset voltage drift are the -00 ones used in non-critical designs.  They actually graded them for 40 volt operation as well instead of buying the 741/A/M version.

One of my PS-503s has a pair of 301A operational amplifiers for U55 and U155 but they and one of the 741s are in sockets so I think someone who knew what they were doing repaired it in the past.  They added the compensation capacitors needed by the 301As to the back of the printed circuit board.

Yeah, looking at it kind of intimidates me a little.  I'm sure everything they did had a purpose at the time and I'm sure there are modern designs that run circles around it but I think simple would be better for me until I learn more.  What I really want to learn is the analog side of things.  When I was a kid and used to take things apart, the electronics always intrigued me but I never really did anything with it.  Up until a few years ago I even had a few of those big assed aluminum variable capacitors.  Digital wasn't here yet and I'd like to start with much less than a microcontroller.

Some multimeters have a pretty fast peak detect mode.

Any storage oscilloscope is handy in cases like these.  DSOs are particularly useful for diagnosing startup problems and even an old and low performance unit can handle something like this.  I very rarely need to use anything better than a Tektronix 2230 which is the oldest DSO that I can honestly recommend.

I would never recommend a USB DSO unless it had a unique and useful feature like the ones from CircuitGear which support low frequency vector network analysis.  That is particularly useful for designing high performance power supplies.

The scope I have is ancient even by ancient standards.  Behold, the Conar 255.  I've been keeping an eye out on Ebay but things go fast there and the resellers are a big problem.  I figure if I'm going to spend over $200 I might as well wait and keep saving to get a DS1074Z.  Every now and then I check craigslist but there never seems to be anything there which is either odd or not considering I live fairly close to Boston.

[liquibyte]

I think I've found a solution that may help.  I've seen the circuit in action on a video and it works but I simplified it a bit for my needs.

[David Hess]

The output should never rise above the set point when power is applied or removed.  I am inclined to believe that there is more than one problem but the lack of any regulation or protection on the supply to the operational amplifiers I consider a design flaw.

That's kind of what I was thinking.  The design has never really worked from what I can tell.  Reading the lengthy thread(s) on the design gives people that have had success building the thing but I don't think anyone has really tested it until very recently.  One of the members over there just recently got a DSO and tried to solve the transient issue with the addition of the voltage regulator but I think he only tested within a certain range and reported it solved.  It wasn't.  I'm about ready to give up on this one for awhile and see if I can get something better going with a bit more research.

Usually these problems are caused by improper startup and shutdown sequencing but linear power supplies are generally simple enough not to require complex solutions.

Any word on a vetted design that would be relatively simple that I could perhaps use the two 28V 4.3A transformers with?  I was shooting for 30V @ 3A but I also just got, in addition to the 301A's, OP07's and TL071's, a single primary, dual secondary transformer that can do 36V @ 1A series, 18V @ 2A parallel or 2X18V @ 1A independently.  I was thinking 0-18V @ 1A positive and negative until I can figure out this other issue, I just need a proven design that won't set me back too much.

The only kit power supply I have is 723 based and 20+ years old.  Since then I have usually designed and built my own as needed.

I did a quick search online and did not find an designs or kits I would recommend.  Variable constant voltage constant current designs tend to be as complex as the comparable Tektronix or HP implementations.

The PS-503 design is pretty dense and unconventional in some respects even if you ignore the complexity added by being a dual output design with tracking.  There are changes I would make like having the output amplifiers operate with a fixed gain and varying the output from the reference because it would make the frequency response more predictable.

...

Yeah, looking at it kind of intimidates me a little.  I'm sure everything they did had a purpose at the time and I'm sure there are modern designs that run circles around it but I think simple would be better for me until I learn more.

The Tektronix and HP/Harrison designers certainly knew what they were doing and were thorough.  HP bought Harrison Laboratory Company in 1962.

Any performance limitations are more from application limits then an old design.  Modern designs intended for the same applications are not going to perform any better at least in a practical sense.

The PS-503 has simple zener shunt and series regulators to protect the control circuits from high voltages.  It has other design features which look like they were intended to prevent startup and shutdown glitches but the manual only discusses one of them.

The similar HP/Harrison 6236 which is a more straightforward design has its own protection circuits.  From its manual:

4-11 The turn-on/turn-off control circuit prevents output
transients when the supply is turned on or off. It does this by
delaying the application of certain bias and reference voltages
at turn-on and removing them shortly after turn-off.

What I really want to learn is the analog side of things.  When I was a kid and used to take things apart, the electronics always intrigued me but I never really did anything with it.  Up until a few years ago I even had a few of those big assed aluminum variable capacitors.  Digital wasn't here yet and I'd like to start with much less than a microcontroller.

Even complex power supplies are simple enough that they are a great place to start learning electronics.

Some multimeters have a pretty fast peak detect mode.

Any storage oscilloscope is handy in cases like these.  DSOs are particularly useful for diagnosing startup problems and even an old and low performance unit can handle something like this.  I very rarely need to use anything better than a Tektronix 2230 which is the oldest DSO that I can honestly recommend.

I would never recommend a USB DSO unless it had a unique and useful feature like the ones from CircuitGear which support low frequency vector network analysis.  That is particularly useful for designing high performance power supplies.

The scope I have is ancient even by ancient standards.  Behold, the Conar 255.

That is very old.  Even the lowest end of the inexpensive used analog oscilloscopes from Tektronix and others would be better but you need some type of storage to see glitches easily.

I've been keeping an eye out on Ebay but things go fast there and the resellers are a big problem.  I figure if I'm going to spend over $200 I might as well wait and keep saving to get a DS1074Z.

The 2230 that I have I picked up on Ebay for about $100 but I would not recommend one for a beginner unless you have the option to return it; they are somewhat tricky to repair and maintain and are getting old.  The 2232 is a lot better but tends to go for $300.

I have not used one but the CircuitGear USB DSOs I mentioned are interesting given their low cost.  They are not particularly fast but they are certainly fast enough for power supply design and look like they were intended for a student environment.

The least expensive Rigol I would consider is the DS1074Z.  I would absolutely avoid at least their B, CA, D, and E series oscilloscopes because they lack peak detection as I understand the term according to Rigol.  The documentation for their Z and 2000A series oscilloscopes is not clear on the matter either and naturally nobody who has reviewed them has tested for what exactly they do support so I am dubious about them as well which just leads me to conclude that Rigol makes junk.

How can they design a modern DSO which supports peak detection or envelope detection but not both and then screw up the documentation repeatedly by confusing the two?  I complained about this to them years ago trying to get answers; at first they did not understand the difference and then they lied about it.

Every now and then I check craigslist but there never seems to be anything there which is either odd or not considering I live fairly close to Boston.

I am near St. Louis and see good analog and digital storage oscilloscopes show up on Craigslist about once or twice a year.  I picked up a Tektronix 547 that way but missed a Tektronix 2232.  I saw something like a Hitachi VC-6145 once and a Leader analog/digital CRT DSO but I prefer more common instruments with good service documentation.

Boston should have a larger market then St. Louis for used test equipment because of the technical industries in the area.  You might want to check out the MIT swap meet.

I think I've found a solution that may help.  I've seen the circuit in action on a video and it works but I simplified it a bit for my needs.

That just looks like a soft start circuit.  Power supplies generally lack soft start circuits except for their current limit and small output capacitor.

[liquibyte]

Usually these problems are caused by improper startup and shutdown sequencing but linear power supplies are generally simple enough not to require complex solutions.

What's kicking my butt about it is that the transformers are only 180VA and the caps are 15000uF.  I've read that inrush is an issue with toroids or transformers around 500VA.  I would think that the circuit could hold the voltage and current based on the pot settings.

The only kit power supply I have is 723 based and 20+ years old.  Since then I have usually designed and built my own as needed.

I did a quick search online and did not find an designs or kits I would recommend.  Variable constant voltage constant current designs tend to be as complex as the comparable Tektronix or HP implementations.

The Tektronix and HP/Harrison designers certainly knew what they were doing and were thorough.  HP bought Harrison Laboratory Company in 1962.

Any performance limitations are more from application limits then an old design.  Modern designs intended for the same applications are not going to perform any better at least in a practical sense.

The PS-503 has simple zener shunt and series regulators to protect the control circuits from high voltages.  It has other design features which look like they were intended to prevent startup and shutdown glitches but the manual only discusses one of them.

The similar HP/Harrison 6236 which is a more straightforward design has its own protection circuits.  From its manual:

4-11 The turn-on/turn-off control circuit prevents output
transients when the supply is turned on or off. It does this by
delaying the application of certain bias and reference voltages
at turn-on and removing them shortly after turn-off.

Even complex power supplies are simple enough that they are a great place to start learning electronics.

I was impressed enough with the service manual on the 6236B that I printed it out and went looking on Ebay.  I think I may be able to live with what I have until I get a better scope and then try and pick up a servicable (by me at least) design or two.

That is very old.  Even the lowest end of the inexpensive used analog oscilloscopes from Tektronix and others would be better but you need some type of storage to see glitches easily.

The 2230 that I have I picked up on Ebay for about $100 but I would not recommend one for a beginner unless you have the option to return it; they are somewhat tricky to repair and maintain and are getting old.  The 2232 is a lot better but tends to go for $300.

I have not used one but the CircuitGear USB DSOs I mentioned are interesting given their low cost.  They are not particularly fast but they are certainly fast enough for power supply design and look like they were intended for a student environment.

The least expensive Rigol I would consider is the DS1074Z.  I would absolutely avoid at least their B, CA, D, and E series oscilloscopes because they lack peak detection as I understand the term according to Rigol.  The documentation for their Z and 2000A series oscilloscopes is not clear on the matter either and naturally nobody who has reviewed them has tested for what exactly they do support so I am dubious about them as well which just leads me to conclude that Rigol makes junk.

How can they design a modern DSO which supports peak detection or envelope detection but not both and then screw up the documentation repeatedly by confusing the two?  I complained about this to them years ago trying to get answers; at first they did not understand the difference and then they lied about it.

I've heard good and bad things about Rigol but I don't really know enough to make an informed decision.  I think the least expensive I'd go for is the DS1074Z too, at least from what I've read.  As far as Tek scopes go, from what I understand, you can get quite a bit of unobtainium down that road.

I am near St. Louis and see good analog and digital storage oscilloscopes show up on Craigslist about once or twice a year.  I picked up a Tektronix 547 that way but missed a Tektronix 2232.  I saw something like a Hitachi VC-6145 once and a Leader analog/digital CRT DSO but I prefer more common instruments with good service documentation.

Boston should have a larger market then St. Louis for used test equipment because of the technical industries in the area.  You might want to check out the MIT swap meet.

The MIT swap meet is coming up on the 21st.  I just talked with the wife and she's even going to let me go by myself.  For what I have I don't mind too much.  I can at least see something which is more than I've had in the past.  Maybe I'll be able to find a few things there that would make life a lot easier.  What kinds of things would you recommend I be on the lookout for scope and power supply wise?  These are the two things on my wants list but I doubt I'll have more than $100 total.  I won't count on it but I'm going to go and see what I can score.

That just looks like a soft start circuit.  Power supplies generally lack soft start circuits except for their current limit and small output capacitor.

It was my understanding that this would allow the filter caps to charge at a slower rate and therefore take the spike out of the equation, or enough of it to make a huge difference before full power is applied to the power supplies.  I'm also assuming that these are rather prevalent in audio amps of a certain power.  I actually have two 10 ohm 100W aluminum resistors that I was going to use and then adjust the time before the circuit kicks the relay to right at the point where the spike disappears. 

[David Hess]

Usually these problems are caused by improper startup and shutdown sequencing but linear power supplies are generally simple enough not to require complex solutions.

What's kicking my butt about it is that the transformers are only 180VA and the caps are 15000uF.  I've read that inrush is an issue with toroids or transformers around 500VA.  I would think that the circuit could hold the voltage and current based on the pot settings.

It does not make a sense to me either.

As far as Tek scopes go, from what I understand, you can get quite a bit of unobtainium down that road.

Custom parts are not usually a big problem because parts donors are common.

What kinds of things would you recommend I be on the lookout for scope and power supply wise?  These are the two things on my wants list but I doubt I'll have more than $100 total.  I won't count on it but I'm going to go and see what I can score.

There are not many options for old Tektronix DSOs worth getting and I would not recommend any of them to a beginner unless they are known to work.  Any old HP power supply should be repairable assuming there is service documentation available for it.

That just looks like a soft start circuit.  Power supplies generally lack soft start circuits except for their current limit and small output capacitor.

It was my understanding that this would allow the filter caps to charge at a slower rate and therefore take the spike out of the equation, or enough of it to make a huge difference before full power is applied to the power supplies.  I'm also assuming that these are rather prevalent in audio amps of a certain power.  I actually have two 10 ohm 100W aluminum resistors that I was going to use and then adjust the time before the circuit kicks the relay to right at the point where the spike disappears.

The problem should not be occurring whether there is a surge on the input or not.

[liquibyte]

I know it's one thing to help in a forum and another in real life but I could send you an unpopulated board or two if you'd like to have a look.  I can populate it with the more common stuff but I figure that wouldn't matter much in the grand scheme of things.  I'm lost with this not behaving properly and I'd like someone that knows what they're doing look at this on a scope.  All of the parts are common stuff except for the output cap I used and maybe the 2141's.  It's ok if you don't want to but I figured I'd ask.  If I could figure this out, I think it would make a pretty good beginner project although it didn't work out for me as it is.

[David Hess]

Before we consider going that far, I have some questions.  Where is the input bridge rectifier and input capacitor C1 in that photo?  What value did you use for C1?  Did you calculate it?

One thing which can cause a voltage surge across the input capacitor is if it is too small although I have never run across that problem myself.

[liquibyte]

Before we consider going that far, I some questions.  Where is the input bridge rectifier and input capacitor C1 in that photo?  What value did you use for C1?  Did you calculate it?

One thing which can cause a voltage surge across the input capacitor is if it is too small although I have never run across that problem myself.

I've got the rectifier and filter cap off board with the transformer.  I also went a little unusual with the BD139's being off board with the 2N3055's.  Links are the datasheets if you need to look at them.  See the pic for the arrangement of everything.  It's messy as far as the wiring goes but I've been slowly cleaning it up and organizing a bit better as I've gotten things where they're going to end up eventually.  For instance, the BD139's are not upside down anymore though they are still on the beefy heatsinks you see in the image.

C1 is 15,000uF because Digikey didn't have a 12,000uF in stock at the time so, if anything, it's larger than was originally calculated.  I didn't calculate the value, I just used the values off the parts list for the schematic with the exception of the filter cap being larger, the other caps upped in voltage to 100V, the transformer being 4.6A instead of 4.3A, and the rectifier being 15A instead of 6-10A.  The output cap C7 is a polyester film 10uF @ 63V because it was recommended to be able to adjust down to true zero.

[David Hess]

The layout makes me wonder if leaking magnetic flux from the transformers is getting into the wiring going to the output transistors or something else.  Leaking magnetic flux could go along with the surge at turn on across the input capacitor but at 12,000 or 15,000 microfarads, it should be large enough to prevent that problem.

But this would not explain the surge problem reported from others with this design so if there is an issue, it is in addition to a problem with the circuit.

Having long wiring runs to the pass transistors may cause problems.  Low value resistors, like 4.7 to 47 ohms depending on the base current, in series with the bases at the transistor sockets could help.  This could especially be a problem with the BD139 because they are pretty fast.

[liquibyte]

The layout makes me wonder if leaking magnetic flux from the transformers is getting into the wiring going to the output transistors or something else.  Leaking magnetic flux could go along with the surge at turn on across the input capacitor but at 12,000 or 15,000 microfarads, it should be large enough to prevent that problem.

But this would not explain the surge problem reported from others with this design so if there is an issue, it is in addition to a problem with the circuit.

Having long wiring runs to the pass transistors may cause problems.  Low value resistors, like 4.7 to 47 ohms depending on the base current, in series with the bases at the transistor sockets could help.  This could especially be a problem with the BD139 because they are pretty fast.

I don't think it could actually be either issue.  I've got a separate board being powered from the second transformer with an on board rectifier, the filter caps and the pass transistor sitting right next to it about three feet away from the transformer and I'm getting the exact same results.  I think I need to start probing around the circuit to see if I can find anything odd going on, I'm just not sure what to be looking for.  Like I said, the wife's letting me go to that MIT thing so maybe I'll come back with a better scope.

[David Hess]

Could the measurement of the surge voltage across the input capacitor and the output at turn-on be in error?  An input surge that large does not make any sense unless the input capacitance is low.

[liquibyte]

Could the measurement of the surge voltage across the input capacitor and the output at turn-on be in error?  An input surge that large does not make any sense unless the input capacitance is low.

At the cap bank, the stabilized output is 43.7 volts and at the power supply output 30 volts (calibrated).  If I allow the caps to drain to a low value, say <5 volts, at turn on I get a consistent 61.2 volts surge at the caps and 37 volts at output (output drained to under 200 millivolts).  So we have a 17.5 volt spike at the caps and a 7 volt spike at the output.  I also tested the output of the transformers independently with four 1N4007 diodes in a full bridge configuration and two 2200uF caps in parallel and I get a 63 volt spike, so this phenomenon is evident outside the actual power supply circuit.  The caps must be drained to see this however.

I can only see this briefly on the meter but the scope is more evident in its display.  I've also measured the AC right at the connections where it comes into the board and I see no surge of any kind with either the meter or the scope which to be honest may not be fast enough to detect anything.  What's happening is doing so at and after the filter cap(s) and I don't have the best equipment to be testing this stuff but they do agree as to the results so I don't think what is happening is an error.  I don't have a second meter to test with and I obviously don't have another scope or even a second channel so until I get something better I'm stuck.  Everything I've read says that with a 130VA transformer this shouldn't be an issue, or be negligible at best.  I guess I'm just special.

[sunnyhighway]

When I see a problem I tend to break the problem down in the smallest things possible and in the process allow for oversimplifying.
Then I look for the simplest solution. Procedural solutions are allowed too. Then look for simplest resolution for the disadvantage(s) of the simplest solution.

Problem: When turned on, a higher voltage than the set voltage is briefly outputted to the terminals, possibly killing or damaging the connected circuit.

Possible solutions:
#1: Start the PSU always with 0V at the output terminals. Wait. Slowly increment the output voltage.
#2: Start the PSU always with the output terminals disconnected. Wait. Connect the circuit to the output terminals.

Both solutions wil work as a procedural solution. But the disadvantage is that procedures tend to be forgotten or skipped.

Final solution:
Implement one (or both) of the procedural solutions using electronics or mechanics in such way it cannot be forgotten.

The mechanical solution for automating procedure #2 could for example be a switch which (when off) covers one or both of the output terminals.

I'll leave other possible solutions as an exercise for the reader.

[liquibyte]

Boston should have a larger market then St. Louis for used test equipment because of the technical industries in the area.  You might want to check out the MIT swap meet.

I want to thank you for pointing out the MIT thing.  I scored a perfectly working 465B with probes and manual for $65.  I may never fix the issues with this design but I think I'll be better equipped to see what's going on.

[David Hess]

The Tektronix 465B and that whole series are really nice oscilloscopes.  The somewhat related 464, 466, and 468 have storage in one form or another which would help in tracking down a voltage surge problem but the 465B is more reliable and certainly better than what you were using previously.

[paul18fr]

Hi

My following remark may not be relevant, but can we imagine than the initial spikes come from the transformer ?

Would a toroidal transformer (instead of a classical one) fix the problem ? (I'm thinking about it after some readings)

Paul

[liquibyte]

The Tektronix 465B and that whole series are really nice oscilloscopes.  The somewhat related 464, 466, and 468 have storage in one form or another which would help in tracking down a voltage surge problem but the 465B is more reliable and certainly better than what you were using previously.

There was a 7000 series there with a full set of plugins but I didn't have the $300-400 to negotiate on price with (looked very much like this).  There were also others and in each case I asked about price and the $300-400 range seemed to be ubiquitous.  I was limited to about $160 and think I really lucked out considering.  The guy that was selling it had it on the ground kind of hidden away from his table but I knew what kind of things I was after and spotted it almost immediately.  I asked him if it worked and he said it shows a nice bright trace.  I asked him how much and he said $65 and I said sold.  I knew if it showed a trace but was uber cheap I might have to fix something so I jumped at it.  I got it home and hooked it up and there was absolutely nothing wrong with it.  As a matter of fact, it's already come in handy due to the dual channels and better features than what I had.  The power supply that I had boards made for is smoother than the one with the voltage regulator which is showing a distinct oscillation at the output.  Either way, I feel like a kid a Christmas.

There are the usual scuff marks from use but no dents.  All of the knobs are there and unbent or cracked or broken or missing pieces and act like they're new.  The feet on the back where the cord wraps around are all broken but that's literally the only fault I can find.  I tried a cursory search and couldn't find any for sale using the Tek part number so I'll have to keep my eyes open for replacements.  I'll be saving my lunch money over time and probably get a 1074Z eventually so I have something that can do storage.

Hi

My following remark may not be relevant, but can we imagine than the initial spikes come from the transformer ?

Would a toroidal transformer (instead of a classical one) fix the problem ? (I'm thinking about it after some readings)

Paul

If anything, I would think a toroidal transformer might make things a tad worse due to the nature of the way they work.  Inrush is happening with the filter caps charging and is always going to be an issue with this supply but there are things we can do about it.  The concern that I have is that the power supply circuit itself seems to lose all control over the voltage at these conditions and the designer(s) probably haven't addressed these situations in the design from the start.  I don't know enough yet to say one way or the other what the eventual solution will be but I am working on it.  Money is not something I have a lot of so it may take more time than if someone that does this for a living was working on it.  I think that in that case though the design would have been scrapped in favor of something else.  Have no fear though, I'm fully vested and refuse to give up and I feel that if a solution exists I'll figure it out.

[David Hess]

There was a 7000 series there with a full set of plugins but I didn't have the $300-400 to negotiate on price with (looked very much like this).  There were also others and in each case I asked about price and the $300-400 range seemed to be ubiquitous.

The 7000 series are generally older and more difficult to maintain so it is probably just as well.  The only significant advantage they would provide over a 465B is 4 channel support and the ability to use specialized vertical inputs.

It would not be my first choice but the 7854 like you linked can operate as a digital storage oscilloscope but not in the way most people would expect because it is not real time capable.  With the right timebase plugin, a 7B87, it can capture single shot events and has pretrigger capability which is exactly what you need to look at what is going on to cause the surge voltage when power is applied.  It can only capture one channel at a time though which would be a significant disadvantage in this case.

There are the usual scuff marks from use but no dents.  All of the knobs are there and unbent or cracked or broken or missing pieces and act like they're new.  The feet on the back where the cord wraps around are all broken but that's literally the only fault I can find.  I tried a cursory search and couldn't find any for sale using the Tek part number so I'll have to keep my eyes open for replacements.  I'll be saving my lunch money over time and probably get a 1074Z eventually so I have something that can do storage.

Water with a little bit if dish soap and a rag is a good surface cleaner.  The front panel controls are easy enough to remove so you can get under them.

The feet are a known issue with these oscilloscopes.  I know at least one person has used a 3D printer to produce very workable replacements and was selling them as sets.

My following remark may not be relevant, but can we imagine than the initial spikes come from the transformer ?

Would a toroidal transformer (instead of a classical one) fix the problem ? (I'm thinking about it after some readings)

The voltage surge at the input is one problem but that by itself should not cause the regulator to behave the way it is so there are at least two problems.

[liquibyte]

The 7000 series are generally older and more difficult to maintain so it is probably just as well.  The only significant advantage they would provide over a 465B is 4 channel support and the ability to use specialized vertical inputs.

It would not be my first choice but the 7854 like you linked can operate as a digital storage oscilloscope but not in the way most people would expect because it is not real time capable.  With the right timebase plugin, a 7B87, it can capture single shot events and has pretrigger capability which is exactly what you need to look at what is going on to cause the surge voltage when power is applied.  It can only capture one channel at a time though which would be a significant disadvantage in this case.

I'm very happy to have stumbled across what I did, the quality of the signal is so much better.  I have no doubt that even if I had a DSO, I'd probably be using this for day to day use.  It came with two 10X probes but I also have a set I got on Ebay that have the 1X - 10X switch.  I've compensated both sets of probes so I now know how to do that.  All I can say is what a marvellous invention.

I've been doing a bit of reading and I think I may be able to use Ch.1 out to capture the event on my computer.  I'll need to get a connector and cable made up for it and it's not ideal but is better than nothing.

Water with a little bit if dish soap and a rag is a good surface cleaner.  The front panel controls are easy enough to remove so you can get under them.

The feet are a known issue with these oscilloscopes.  I know at least one person has used a 3D printer to produce very workable replacements and was selling them as sets.

The case was a bit dirty but I did clean it with 401 countertop cleaner and it brightened up nicely.  I'm going to hold off on the front panel because, to be honest, it's fairly clean.  I'll wait to do that part until I get a chance to open it up and assess the innards.

My following remark may not be relevant, but can we imagine than the initial spikes come from the transformer ?

Would a toroidal transformer (instead of a classical one) fix the problem ? (I'm thinking about it after some readings)

The voltage surge at the input is one problem but that by itself should not cause the regulator to behave the way it is so there are at least two problems.

There was an idea by redwire over at the other forum that the circuit in the attached image might help.  I don't have the TIP141 to test this out with and I have to wait to place another order at Digikey until the 3rd of next month.  I've got a relay sitting there too, just in case.

[David Hess]

The 7000 series are generally older and more difficult to maintain so it is probably just as well.  The only significant advantage they would provide over a 465B is 4 channel support and the ability to use specialized vertical inputs.

I'm very happy to have stumbled across what I did, the quality of the signal is so much better.

Brightness and clarity of the trace is often under appreciated.  Higher bandwidth analog oscilloscopes generally have higher acceleration voltages making them brighter and clearer even on low bandwidth signals.  The extra brightness is also convenient when looking at low repetition rate or single shot signals.

On a DSO this is irreverent except insofar as a CRT with high acceleration voltage allows for a higher resolution display which that 7854 takes advantage of by displaying a full 10 bits horizontally and vertically.  It has a 10 bit digitizer as well.

I have no doubt that even if I had a DSO, I'd probably be using this for day to day use.  It came with two 10X probes but I also have a set I got on Ebay that have the 1X - 10X switch.  I've compensated both sets of probes so I now know how to do that.  All I can say is what a marvellous invention.

Here is a video of Jim Williams talking about his Minuteman guidance computer and repairing a 465B owned by Tom Osborne of HP calculator fame:



When I went DSO hunting a couple years ago and before the Z series were available from Rigol, I decided to hunt down a Tektronix 2230 which is newer than the 465B by at least 3 years.  It is a combination analog and digital storage oscilloscope and the descendant of the Tektronix 468.  Its major advantage over the non-Z Rigols other than cost is that it supports real peak detection which largely alleviates aliasing concerns at slow sweep speeds which was a major consideration.

I've been doing a bit of reading and I think I may be able to use Ch.1 out to capture the event on my computer.  I'll need to get a connector and cable made up for it and it's not ideal but is better than nothing.

That could work and is pretty close to what the external vertical output was intended for.

Water with a little bit if dish soap and a rag is a good surface cleaner.  The front panel controls are easy enough to remove so you can get under them.

The feet are a known issue with these oscilloscopes.  I know at least one person has used a 3D printer to produce very workable replacements and was selling them as sets.

The case was a bit dirty but I did clean it with 401 countertop cleaner and it brightened up nicely.  I'm going to hold off on the front panel because, to be honest, it's fairly clean.  I'll wait to do that part until I get a chance to open it up and assess the innards.

I have heard reports before of 401 cleaner removing some types of labeling from instruments.  Of course soap and water will remove decals if not used carefully.

There was an idea by redwire over at the other forum that the circuit in the attached image might help.  I don't have the TIP141 to test this out with and I have to wait to place another order at Digikey until the 3rd of next month.  I've got a relay sitting there too, just in case.

That could work but it should not be needed if the supply voltage to the operational amplifiers is regulated.

There are a couple of possibilities but I think we narrowed it down to just one:

1. The supply voltage to the amplifiers was exceeding their absolute maximum during the surge.  We fixed this with a simple zener shunt regulator which should have been present anyway.

2. The input voltage to the pass elements might be high enough to couple through via breakdown.  I do not think this is happening because BD139 is an 80 volt transistor and the pass transistors are 2N3055s.  Maybe the 2N3055s are defective or counterfeit.

3. The output U2 showed the same glitch.  That needs to be tracked down by looking at if the same glitch is showing up earlier in the signal chain at the non-inverting input of U2.  If it is *not* there, then maybe the output is being dragged up which can be tested by disconnecting it from Q2 and shorting it to the other end of R12 which removes the pass transistors from the circuit.