Test Equipment Anonymous (TEA) group therapy thread

[bd139]

Quick question for everyone: what’s the cheapest and easiest way to measure frequency at 10GHz to 1Mhz precision? Approx available power 1mW.

Someone has just offered me a box of 10GHz stuff and I’m working out if it’s worth it as it involves quite a drive.

Edit: can that idea. I can’t be arsed and someone closer wants it.

[tggzzz]

Quick question for everyone: what’s the cheapest and easiest way to measure frequency at 10GHz to 1Mhz precision? Approx available power 1mW.

Someone has just offered me a box of 10GHz stuff and I’m working out if it’s worth it as it involves quite a drive.

Edit: can that idea. I can’t be arsed and someone closer wants it.

So, 1 part in 10⁴, which is undemanding.

Obvious candidates would be a frequency counter (optionally with prescaler but check sensitivity) or a spectrum analyser (optionally fed from a known 10MHz source). I have more of the latter lying around

[nixiefreqq]

Quick question for everyone: what’s the cheapest and easiest way to measure frequency at 10GHz to 1Mhz precision? Approx available power 1mW.

Someone has just offered me a box of 10GHz stuff and I’m working out if it’s worth it as it involves quite a drive.

Edit: can that idea. I can’t be arsed and someone closer wants it.

never used one myself.  but a buddy of mine swears by the old gumball units.

https://www.ebay.com/itm/HP-537A-AGILENT-537A-COAXIAL-FREQUENCY-METER-3-7-TO-12-4-GHz/333280868391?hash=item4d99167827:g:rIQAAOSwP2ddPyor

[AVGresponding]

Spent a couple of hours this morning making myself some Kelvin leads:



Used butyl 2.5mm sq flex and a crap-ton of heatshrink   
Neutral and earth together for the sense, and line for the resistance measure.
The reason I used banana plugs instead of soldering direct to the clips is so I can use my fiendishly sharp RS probes for chasing down shorts as well as the clips.
I realise I sacrifice a tiny bit of performance, but it's not much:



That's with them shorted and after a brief warm up. After a couple of hours the offset has dropped to 2 milliohms.


It's impressive the lengths some will go to in order to keep us informed:

[med6753]

More calibrations today. Fluke 8600A. This one did not go well and in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.



So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.   

[bd139]

That doesn't surprise me with the 8600A. I had trouble with mine. One of the reed relays was duff which caused it to jump range and then it drifted slightly as well.  At this point, and this is probably controversial, I think was a 3466A is a better meter than the 8600A was, assuming some fucknuckle hasn't put an IC in it the wrong way round  . Then again sample size is one each here.

[nixiefreqq]

More calibrations today. Fluke 8600A. This one did not go well and I in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.

So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.

so yer 40 year old meter is off by 4 mV when measuring 10V?

glad to see you did not fall down the rabbit hole.

edit (also like my 3466a.  and it is fine for its main use........to see if something is totally dorked.  RCH measurements rarely matter to me.)

edit 2 RCH voltages don't matter to me.  measuring frequency is a totally different matter.  anything less that 1 part in ten to the 9th makes my teeth itch when it comes to frequency.

[worsthorse]

More calibrations today. Fluke 8600A. This one did not go well and I in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.

So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.

hmmmm.... that looks like a rabbit hole, for sure. mebbe better to declare victory and spend that time on something, well, more damaged, in the queue?   

btw, I just went off and dug up your voltage reference schematics. Thanks for posting them. I am going to check the 8840A with my low voltage reference and the 0.1% 10 to 10M ohm board this morning.  Running it through the Fluke calibration procedure, though, is going to be problematic as the procedure is automated and seems to require specific Fluke calibration gear. I am going to cozy up with the manual today and see if I can fake it out.

btw two, the repaired Fluke is now passing all the self tests. I am going to let it burn in with the case off and look at it again in a couple of hours, before putting the case on and letting it run for the day.

[med6753]

Continuing with more calibrations. Next up is the Mastech MS8040 4.5 digit. People write these off as cheap Chinese crap but this DMM does have reasonable accuracy and it remains in calibration. No adjustments were required. And while I don't use it all that much it does have some desirable features such as temperature (both in C and F) and a reasonably accurate capacitance function.




Still have more to go. Fluke 8010A and Fluke 8810A.

[med6753]

More calibrations today. Fluke 8600A. This one did not go well and I in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.

So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.

so yer 40 year old meter is off by 4 mV when measuring 10V?

glad to see you did not fall down the rabbit hole.

edit (also like my 3466a.  and it is fine for its main use........to see if something is totally dorked.  RCH measurements rarely matter to me.)

edit 2 RCH voltages don't matter to me.  measuring frequency is a totally different matter.  anything less that 1 part in ten to the 9th makes my teeth itch when it comes to frequency.

Well, when your Chinese Mastech reads it exactly at 9.9970V it becomes a matter of principle. So I will be visiting that rabbit hole but I may never find the bottom. 

[nixiefreqq]

More calibrations today. Fluke 8600A. This one did not go well and I in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.

So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.

so yer 40 year old meter is off by 4 mV when measuring 10V?

glad to see you did not fall down the rabbit hole.

edit (also like my 3466a.  and it is fine for its main use........to see if something is totally dorked.  RCH measurements rarely matter to me.)

edit 2 RCH voltages don't matter to me.  measuring frequency is a totally different matter.  anything less that 1 part in ten to the 9th makes my teeth itch when it comes to frequency.

Well, when your Chinese Mastech reads it exactly at 9.9970V it becomes a matter of principle. So I will be visiting that rabbit hole but I may never find the bottom. 

damn.  we are gonna' miss ya' med.

say hi to alice for us.

[AVGresponding]

Continuing with more calibrations. Next up is the Mastech MS8040 4.5 digit. People write these off as cheap Chinese crap but this DMM does have reasonable accuracy and it remains in calibration. No adjustments were required. And while I don't use it all that much it does have some desirable features such as temperature (both in C and F) and a reasonably accurate capacitance function.

Not never!

My Mastech clamp meter has ridiculous DCA accuracy, far better than the ~£20 price tag warrants, matches readings obtained with a 1% shunt, so I will give them a huge benefit of the doubt (given that I dislike Chinesium products generally).


EDIT: The Fluke has had about 6 hours to warm up now, and the home-made kelvin leads show a 1-2 milliohm offset. That's good enough for the UK as well as Australia!

[bd139]

Stupid idea incoming. Just ordered parts for the next test gear related project which is going to be fun. Building a PLL unit for the HP3312 (or any other FG). Sync output -> PLL -> VCO input. Going to see if I can lock it to 1KHz steps. Pretty simple. One each HC4059 divider, HC4046 pll, HC4060 ref osc, LMC6482 and some RC loop filter parts. Because they're all bloody SMD now, rather than actually do a prototype I've gone straight to board and ordered that as well . Will it work? Probably not 

[med6753]

Stupid idea incoming. Just ordered parts for the next test gear related project which is going to be fun. Building a PLL unit for the HP3312 (or any other FG). Sync output -> PLL -> VCO input. Going to see if I can lock it to 1KHz steps. Pretty simple. One each HC4059 divider, HC4046 pll, HC4060 ref osc, LMC6482 and some RC loop filter parts. Because they're all bloody SMD now, rather than actually do a prototype I've gone straight to board and ordered that as well . Will it work? Probably not 

If it doesn't work you can blow it up.....which you ARE good at. 

[bd139]

  It's all CMOS shit so I'm sure it'll be dead within seconds in my hands anyway 

[Specmaster]

More calibrations today. Fluke 8600A. This one did not go well and in fact stopped the process. It's out of spec.

When I first got this 8600A it had severe drift issues. In the process of attempting to repair I replaced the PSU caps, some tants and poly caps on the main board, the 1MHz oscillator crystal, plus a few other parts. I thought the problem was the zener reference itself but it has tested rock stable. I finally tracked the drift down to contact resistance on some of the range reed relays. So I replaced all of them. That finally fixed the drifting issues. I thought all the issues were fixed but apparently not.

I can calibrate the 190mV, 1.9V, 19V, and 190V and they are repeatable and right on. But if I check other DC voltages with the AD584-M the 2.5V, 5.0V, and 7.5V are OK but applying 9.99691V I get this:

10.001V. Way out. Should be max 9.9970V.



So I stopped the cal for now and this guy is in the repair cue. But I think it's going to be a tough issue to track down may not be worth the effort.   

Jeez you're a hard person to please aint ya? I would consider that being just 4mv after the many years of service, a win personally, come on now are you really going to noticing the differance in practise? 

[med6753]

If it were a lab they would bin it immediately. So why should I accept anything less? 


Besides....some of you are saying "You need 6.5 digits" but at the same time you want me to accept a 4.5 digit DMM that's 4mV out?

Duh....       

[Specmaster]

While that might well be true, a lab has a lot more of disposable green backs then us retired gentlemen have  and then again they are earning a living with them 

Seriously if you're going to be chucking it in a bin, chuck it my way, its most certainly good enough for my work and while you're about it, you might like to chuck your old 8000A as well because that can't match the 8600A   

[med6753]

While that might well be true, a lab has a lot more of disposable green backs then us retired gentlemen have  and then again they are earning a living with them 

Seriously if you're going to be chucking it in a bin, chuck it my way, its most certainly good enough for my work and while you're about it, you might like to chuck your old 8000A as well because that can't match the 8600A   

I didn't say I was going to bin it. I don't give up that easy.

Yes, the 8600A (when working properly) is more accurate than the 8000A. But what the 8000A has over this 8600A is CONSISTENT accuracy across all ranges.

[Neomys Sapiens]

Woohoo just hit the function generator jackpot!



Been after one of them for ages. Was pretty cheap and no knob damage or missing caps as well.

Edit: this is why. These things are top notch engineering even though they are plain on the outside.

Yep, that was my first professional FG. Before there was only a bodge with a 8038 and a project to build something more decent, which never came to fruit. It is still around and working perfectly. Aquired it from HTB in the late 90's.

[med6753]

Yep, that was my first professional FG. Before there was only a bodge with a 8038 and a project to build something more decent, which never came to fruit. It is still around and working perfectly. Aquired it from HTB in the late 90's.

Here's a 8038 IC in action that I built 30 years ago. I use it as a quick scope checker. The outputs are fixed at 5Khz/0.5V p-p. There's also a triangle wave available.

[Specmaster]

While that might well be true, a lab has a lot more of disposable green backs then us retired gentlemen have  and then again they are earning a living with them 

Seriously if you're going to be chucking it in a bin, chuck it my way, its most certainly good enough for my work and while you're about it, you might like to chuck your old 8000A as well because that can't match the 8600A   

I didn't say I was going to bin it. I don't give up that easy.

Yes, the 8600A (when working properly) is more accurate than the 8000A. But what the 8000A has over this 8600A is CONSISTENT accuracy across all ranges.

Thats my point, the 8000A is never at the 10V range going as accurate as the 8600A, even with its 4mv shift and given that (at least in my lab) there seems to much radiated EMF and other noise issues that the last 2 digits seem to be almost constantly in a state of flux that the most stable of the decimal places seems to mainly restricted to the first 2. The last 2 digits, depending on where there are sitting may make the 3rd place OK or it can flip back on forth based on the last 2 going higher then 99 and dropping below 99 can make the 3rd place change up and down by 1 digit 

In most labs where I have ever worked, if the voltage reading was critical, then those labs are normally engineered to be as noise free as possble by have the lab screened and the gear contained within also specially designed. How far down the rabbit hole do you want to go?

[mansaxel]

Holy Fucking ASSCRACKERS!!!  WERE THEY THINKING?!?

mnem
Probably some ignorant BS color scheme generated by some complete backbirth lobbyists because those 4 colors were somehow most profitable to some manufacturer or somesuch wank...

It is the standard EU scheme. It's been like that for at least 30 years, which does not exclude your theories from having influenced it, but that, then, must have happened a long time ago.

[mnementh]

Welp... The lady at the border crossing looked at us and the kids exactly twice, then waved us through. We're on the QEX about 45 minutes from Toronto.

Kosmic, think you should lock up the women & hide the fried chicken, eh...? Hoser-Dwagon is officially in da Great White North!!!

I just hope everything goes as smoothly with the rental property...

[xrunner]

Been busy with a 6 meter beam antenna. I took it down for the winter and I'm fixing it up a bit. The guy that made it for me in the spring had his friend 3D print some clamps for the director and reflector. But as you can see in the picture that person didn't use 100% infill - far less. It's a wonder it held up through all the storms. I have re-printed them (one on the left in the pic) with 100% infill and they are nice and strong now.

The plastic part that supports the driven element is from an old TV antenna. I might as well just re-design it because I have plenty of time, shouldn't be too hard. I don't need all the fancy angles on it - just basically a U shape over the beam support and U shapes going out either side. On top some ridges on either side where the rod sits to prevent it being torqued out of position by the wind (which happened during strong storms). And some screw holes ... done deal.