Bought TTi EL302P off ebay, dead display

[staze]

K, will post back results from local shopping. =)

Guess the big question... I think the local shop carries NTE components... do you know if there is equivalents? I'm checking their cross reference, but not seeing anything for the IRFP250N, or the OP07. =/

Will have to see the actual store stock. Maybe they have some other stuff.

[staze]

Up, local places had nearly none of that stuff... and what they did have was hella expensive. sucks.

Anyway, ordered the stuff off ebay. guessing a couple weeks unless something has changed since I ordered a bunch of 1/4w resistors from china a few weeks back.

For now, I guess when I get the PSU, I'll do basic testing with a halogen bulb for load.

Will post back once parts are here, as well as once I have the PSU in hand. Still need to look at fixing the displays. =)

[IanB]

If you have a serial cable you can control the power supply from a computer using any convenient terminal program. The supply uses a very simple set of commands for control and read back that you can enter by hand.

[amspire]

as for the mosfet , IRFP250N and IRFP064N are good choice for that linear task since they have good SOA ratings in the linear region

I didn't think these parts had any DC SOA specification.

What that means is that these  parts could work fine, or they could fail in linear DC mode. The manufacturer is not giving any opinion, so it is up to you to experiment. It is very likely the designers of these parts did not design them to be suitable for DC linear mode operation.

An example of a device specified for Linear DC operation is the IXYS IXTH24N50L. With the case at 60 degC, it can safely handle something like 5A up to about 35V. It will do 10A up to 20V. The package is 400W at 25 degC so you might be able to get 200W out of the device on a great heatsink.

There are others - IR have devices with linear DC SOA specs - but it is not easy finding them. Look for devices with gate turn on voltages (Gate Source Threshold Voltage at 250uA)  over 3V. The higher the gate turn-on, the more chance the cells will share current evenly allowing the device to handle DC. Also the devices with a higher Rds are more likely to have a DC SOA  rating then the devices with a low Rds (given the same power rating for the two devices).

Richard.

[mariush]

I've built a constant current load using the circuit here:  http://www.sleepyrobot.com/?p=136

But I've used a FQP50N06 : http://www.fairchildsemi.com/ds/FQ/FQP50N06.pdf

It works OK, maxes out at somewhere under 1A if the voltage is below 3-5v or so, above I tested it up to 2-3 amps with my linear power supply (which does max 3a) and it worked great.

My prototyping board soldering skills aren't that great, I've used network cable wire to make connections, soldered the decoupling cap directly on the opamp power pins, works just fine.

[staze]

If you have a serial cable you can control the power supply from a computer using any convenient terminal program. The supply uses a very simple set of commands for control and read back that you can enter by hand.

Yes, saw that. do you know if it just works with a standard serial cable?

It looks like you can just feed it very simple commands followed by an LF and it responds... so that should be easy enough. =) Big reason I bought the supply... worst case, if I can't get the display working, at least I have an easy way to read output... just write something to update on my screen every 20ms or so (according to the manual, you have to wait about 10ms between commands). =)

Thanks IanB!

[staze]

I've built a constant current load using the circuit here:  http://www.sleepyrobot.com/?p=136

I actually saw this design earlier. Great design actually. I bought a 10turn 50k pot just so I could build that design if this one doesn't work. =P

bought duplicates of the opamp and mosfet as well.

guess this is how we build parts bins...

Your soldering skills look fine given it's perfboard. 

Anyway, should be easy enough to pick up those extra parts and breadboard. The previous version should work with a 50k 10turn I'd think. I ordered the couple missing pieces I didn't have real quick, and I'll just pick up the ceramic caps locally... because finding the .1uF is spendy. Guess I could just order from Mouser. Shipping is the killer though on something like that.

[eevblogfan]

hey amspier

I've bought 10 IRFP250N

they indeed don't state the SOA in datasheet , but have a look , they do specified the max current and max voltage , plot it into the current SOA and you'll get the DC SOA ,

no need to spent big $$$ on special mosfet when silly irfp250n can do the task

as for the op , you did well , it's fine any circuit that you've build is okay , the reason I gave you that list was because I built that circuit and it works fine with no oscillation up to few amps . and has -3DB of around 87Khz  ( if modulation needed )

Cheers

[amspire]

they indeed don't state the SOA in datasheet , but have a look , they do specified the max current and max voltage , plot it into the current SOA and you'll get the DC SOA ,

It doesn't work that way unfortunately.

Most power mosfets have no DC safe operating spec because their design does not share current evenly between individual mosfet cells in the mosfet in the linear mode. The current sharing is fine when the mosfet is fully on - that is no problem as the Rds defines the current sharing.

Most power mosfets are not made as a single mosfet device - they are an IC containing a large number of small parallel mosfet devices. The larger the silicon die, the more mosfet cells are in parallel. This is one of the reasons it is easy to scale a design from a low current version up to a high current version - just make the chip bigger and add more cells.

In the linear mode in most power mosfets, the sharing tends to be defined by the gate sensitivity which has a negative temperature sensitivity. What that means is that the hotter an individual cell gets, the more sensitive the gate becomes, so the harder that cell turns on.  So in linear mode, if one cell in the mosfet gets a bit hotter then the other cells, it turns on even harder and gets hotter until thermal runaway occurs shorting the cell and destroying the mosfet. Mosfets that are good at linear mode often have a high enough channel resistance so that it starts to dominate over the gate sensitivity before thermal runaway occurs. Mosfets designed low low Rds switching usually do not have this safety mechanism.

There is no possible way to deduce a mosfet DC SOA from voltage and current specs. Mosfet manufacturers never leave a DC spec of the SOA curves by accident. If there is no DC curve, the designers are not prepared to specify any safe level of DC in linear mode for that device.

As I said, you can try the irfp250n and if you are lucky, it may work fine. It will likely work fine up to a point - perhaps an amp or to. Beyond that it is a gamble. If it blew up and you told IR, they would shrug their shoulders and say "What do you expect - it is not designed for linear DC use". If you were going to make a current load product to sell, you would not touch the irfp250n. You would definite go for a device like the IXYX device I mentioned or the FQP50N06 that mariush mentioned.

[eevblogfan]

I tried the IRFP250N and I can report it held 150W for 24/7 , It may not hold 150W for high MTBF but it may hold okay at the 100W region ( based on my approch to that , it should be SOA max at 110W or so )

please tell me, 30A rating , what is that telling you >?

for me It says that up to 30A the internal arry will shear fine the current , since that's 250V mosfet , it should hold up to 250V peaks at 30A switching , obviously the duty factor of the spikes should be very low , so the AVJ power dissipated will not exceed the 110W I've mentioned , now since I've tried that IRFP250N I can recommend him for that purpose ,

please note BK8500 uses that mosfet for 45W continues maX , I believe you can use this for up to 110 and you should hold good margin so I believe 80~90W continues will be okay for very long MTBF

as per the construction of fets , amspier , I am awear of that , but thank you for bothering writing that long and detaild post , I appreciate that 

Cheers

[staze]

Forgive the n00b question here, but with a 3 post power supply (neg, ground, and pos), if I am hooking up a load, or say, a multimeter, should I be hooking to pos and neg, or pos and ground? If I understand correctly, if I hook from pos and ground, I'd get say +15v (for example). Same as hooking between pos and neg. But if I hooked up between neg and ground, I'd get -15v?

Basically, if I'm not needing anything with a neg rail, I should just hook up between pos and neg, and forget the ground, yes?

[ve7xen]

The positive and negative are the floating supply outputs. The ground is mains earth. If you want a floating supply, ignore the ground lug, but sometimes you will want to tie part of your circuit to earth ground. Sometimes you'll see the negative output strapped to earth. It's not a power supply output though, just a convenient connection to the mains earth pin on the input power cable.

You should get no current between + and ground (unless you've strapped - and ground together, of course).

If you want a +/- supply, you need a second supply. Because they're floating, you can connect them in series and use the centre node as ground.

[staze]

okay, that makes sense. Guess I'm still a bit fuzzy on "floating", but I'm getting there.

Thanks!

[staze]

Sooooo....

Power supply arrived, and looks like the attached photo. =(

Postal system seem to have completely destroyed it. We shall see what the seller and/or ebay do. Doubt it was insured by the seller.

*sigh*

[staze]

That said... anyone got any recomendations for similarly spec'd power supplies for not a whole lot different price (sub-$100, output toggle, 0-20v+, 2 or so amps with either 10turn or encoders?)

  *bummed*

[Monkeh]

.... Insurance or not, that's totally unacceptable.

[IanB]

Sooooo....

Power supply arrived, and looks like the attached photo. =(

Postal system seem to have completely destroyed it. We shall see what the seller and/or ebay do. Doubt it was insured by the seller.

*sigh*

WTF? How was it packed, exactly?

[staze]

it was actually packed pretty well. it was in bubblewrap, then in a good 4 inches all the way around in peanuts.

The box was pretty destroyed too. USPS at its finest I guess.

Sad really. If it turns out I don't have to ship it back, I'm obviously going to try to repair. The front looks like it's in good shape, but I have no idea how the inside looks at this point. =( And of course today is when I started receiving parts for my adjustable load...

Guess I'll be looking at E3610a's...

[ve7xen]

That said... anyone got any recomendations for similarly spec'd power supplies for not a whole lot different price (sub-$100, output toggle, 0-20v+, 2 or so amps with either 10turn or encoders?)

  *bummed*

The Power Designs stuff is ancient, but well designed and stable and can be had for well under $100. I don't think they have output toggles though. Decent examples can be had for under $50. e.g. this. The older HP 62xx series fits the bill too, just be careful you pick a model with constant current and with the output voltage/current combo you need, there are lots of different ones.

If you keep your eyes open you should be able to get an E361x for around $100 too. My E3610A was $85 and that was on the high end of the spread for a guy selling a dozen of them over a week or two. Just keep your eyes open for them or set up e-mail notifications.

Or you could go new, the Mastech stuff is supposed to be decent and some of their models can be had for around $100. http://www.mastechpowersupplies.com/

Also that is REALLY RIDICULOUS. Someone clearly threw it off the loading dock onto concrete. Absolutely not acceptable . Sorry for your loss .

[staze]

Also that is REALLY RIDICULOUS. Someone clearly threw it off the loading dock onto concrete. Absolutely not acceptable . Sorry for your loss .

Think it was machinery. Given how the box looks, I don't see how a fall could have done that. This was crushed.

Sad.

[staze]

So seller fully refunded. Didn't want it back....

Sooo... I've pulled it all apart at this point, and THINK it's salvageable. Innards all look intact except one wire that goes between the RS-232 interface, and the front board. Easy enough to replace.

Going to pound the case back into shape this week(end) and see what I can get. Wonder if TTi would just sell me the casing. =D

[KJDS]

I'm glad I don't ship stuff round the states if that's what happens to them. It's been a long time since I've seen anything that badly damaged and the box on that occasion had a set of lorry tire marks across it. That has suffered a lot more than a drop, it's either fallen off a moving vehicle or been in one that's been crashed, or possibly driven over.

Good luck getting it working.

[grumpydoc]

Agree that it looks crushed, not merely dropped - but, ouch!

Looks like this Tektronix 'scope could have been in the same container!

[free_electron]

Agree that it looks crushed, not merely dropped - but, ouch!

Looks like this Tektronix 'scope could have been in the same container!

Wow that's the best looking tek scope i have ever seen. Must be the salvador dali edition...

[staze]

IanB (or free_electron),

Any chance you could confirm the transformer in yours is wired like this?

Mine still seems to still be wired for 230V, which makes me wonder if that's what causes the display issue. Currently it's wired to the two outside lugs, and then a jumper between the middle ones.

I THOUGHT it would just be unplugging, but it looks like soldering is involved (which is completely fine). I just want to confirm the wiring before doing it. Guessing in 230v config, the middle winding is unused, and in 115v, you are just adding that winding to up the voltage/current on the output side of the transformer.