Temperatures inside Keithley 192 bench meter

[giosif]

Hi,

Not strictly a repair topic (or maybe it is and I am yet to realize it), but I was wondering if it is normal to have the transformer and the smoothing cap for the digital 5V rail in my Keithley 192 meter running at 50+ degrees Celsius (see below picture for the two components and the way I measured the temperature).
Specifically, the transformer runs as hot as 58 degrees and the cap goes up to a little over 51 degrees (both feel hot to the touch).
Otherwise, the meter seems to be working fine.
Could this be related to the fact the transformer supports only 115V and 230V mains (currently set to 230V, of course) while the mains here sit at 240V?

Thanks!

[abdulbadii]

not knowing device
it's likely still acceptable though very rare, for low freq.  usual T i.e. not T for SMPS
it's something serious thing otherwise

[CaptDon]

Are your mains 50Hz?? I suspect the transformer may be approaching saturation at 50hz but would be happier at 60hz. Is the cap simply heating up from the radiated heat of the transformer? Seems like a marginally designed transformer.

[giosif]

Thank you both for the answers!

Yes, mains here are at 50Hz.
About the cap heating up, there is a voltage regulator (U127 for the analog 5V rail) even closer to it than the transformer is and that regulator seems to be getting even hotter (yet to measure the actual temperature).
For the transformer design, I would have thought a company such as Keithley wouldn't have gone with marginal designs (could be wrong, though, and there can always be exceptions).

I would also be curious if someone here, ideally based in a location with 220-240V @ 50 Hz mains, with a working Keithley 192 (they seem to be quite rare, for some reason) is seeing the same thing (mine gets to these temperatures after about 30-45 mins of operation).

[abdulbadii]

Please examine, check the these reasons which most possibly causing it
1. Total load exceed transformer kVA rating.
2. Ambient temperature exceed transformer ratings.
3 Low Power Factor
4. Check the transformer is correctly rated for harmonic load, check for high neutral currents.
5. Transformer’s ventilation openings can be blocked
6. Fan cooled transformers have broken or misaligned fans.
7. Excessive dust can block air vents as result in overheating of transformer..

[The Soulman]

Thank you both for the answers!

Yes, mains here are at 50Hz.
About the cap heating up, there is a voltage regulator (U127 for the analog 5V rail) even closer to it than the transformer is and that regulator seems to be getting even hotter (yet to measure the actual temperature).
For the transformer design, I would have thought a company such as Keithley wouldn't have gone with marginal designs (could be wrong, though, and there can always be exceptions).

I would also be curious if someone here, ideally based in a location with 220-240V @ 50 Hz mains, with a working Keithley 192 (they seem to be quite rare, for some reason) is seeing the same thing (mine gets to these temperatures after about 30-45 mins of operation).

Although my 192 is still a bit of a ongoing project, I've seen similar temperatures.
Especially the 5 volt digital rail has a big voltage drop over the linear regulator, powering the meter thru a small transformer to reduce mains voltage by 24 volt or so does help.

[SeanB]

Small transformers are generally going to run hot, they lack the winding volume to get more thicker wire in, so the resistive losses are high, and limited lamination space means the core runs at higher magnetisation, and thus higher losses. 20C or so above ambient not a problem, you would find most are specified for operation up to 110C surface temperature, with class H insulation allowing 130C deep in the windings.

Capacitor will be conduced heat from the regulator, making life shorter, and you can fix in one of three methods. Replace with 105C version, or put a bigger heatsink on the regulator, so it runs cooler, or just ignore it, and accept the capacitor life will be lower than running at room temperature.

[CaptDon]

Abdul's answer cracks me up!!! Sounds like it was copied word for word from an industrial book concerning 120/240 or 240/480 industrial transformers of 1kva and up. Haven't seen cooling fins and fans on transformers in digital multi-meters lately. Is your Keithley 192 transformer oil cooled?? Perhaps check the heat exchanger for dead birds or rodent nests?
On the serious side, a lot of test equipment with 10va to 25va power transformers really seem to hate 220/240 50hz mains. B.T.W., if your Keithley is oil cooled I have 5 gallons of Inerteen 50% P.C.B. contaminated transformer oil I'll donate to you!!! Sadly, the GG1 locomotive it was used in is long gone!!!

[lowimpedance]

 ....

 It appears to be a design feature of these 'brown box' vintage Keithley's as I have a 196 that also gets rather toasty. Definitely check the filter cap has not dried out, otherwise its a live with situation  .

[giosif]

Thank you all!

Ok, so it appears this is likely "by design" and, if I want to improve things, I'd be looking at replacing the cap and adding a beefier heatsink to the voltage regulator.

[CAL Enable]

Hello giosif,

My first post on this nice forum...

My post is may be a little (too) late, but it happens that I also have a Keithley 192. And I thought, let's have a look at mine.
I made some pictures with an IR-camera and came to the conclusion that, in this case, the transformer doesn't get hot at all.
What get's hot is indeed the voltageregulator and the bridgerectifier even more!

After half an hour powered on with the case closed (ambient temperature 22 C):
Transformer: 33 C
Voltageregulator: 60 C
Bridgerectifier: 78 C(!)
The big cap: 42 C

My 192 is without the option 1910 (= lower powerconsumption) and to be able to messure temperatures, I removed the GPIB-board (option 1923) that's placed right above the power supply part.
May be this information can help you. I shall also try to upload the images.

Kind regards

[giosif]

Hi,

Thank you for taking the measurements!
It is good to have readings to compare against.
And, no, it's not too late. 

For the transformer temperature, though, isn't the part of the transformer that you measured the temperature for reflective?
If so, the measurement from your IR-camera might not be accurate.

[srb1954]

Hi,

Thank you for taking the measurements!
It is good to have readings to compare against.
And, no, it's not too late. 

For the transformer temperature, though, isn't the part of the transformer that you measured the temperature for reflective?
If so, the measurement from your IR-camera might not be accurate.

Quite apart from measuring errors from emissivity variations an IR camera doesn't really tell you much about the more important parameter which is the internal temperature rise of the transformer windings.
 
A better technique for measuring transformer temperature rise is to compare the change of the winding resistance from cold to hot. Using the copper temperature coefficient of 0.39%/C one can calculate the average internal winding temperature from the ratio of the hot to cold resistance.  There will be some small variation of the temperature across the winding so you need to add a safety margin of, say,  another 10C to determine if any part of the winding exceeds the allowable temperature rise appropriate to its insulation class.

[Kleinstein]

The pattern visible in the transformer core part points to poor emissivity. The actual core temperature should be much more uniform. Also the winding insulation (usually plastics and thus usually good emissivity)  looks warmer, but still not too hot. There is nothing wrong with a transformer that does not run excessively hot. Mexico is 60 Hz mains and would thus likely lower transformer loss than 50 Hz.

[CAL Enable]

Because of the comments, in the mean while, I have taken a few new measurements.
In this case all parts stayed within the unit, but it was just possible to measure the temperature of the coils and bridge rectifier.
The unit was powered on for 45 minutes (235V / 50 Hz):

Transformer coils: 43 C
And of course, the internal of the coils shall be warmer.

Iron part of the transformer: 35 C
So just 2 C higher after 15 minutes more, compared with the first measurement.

Bridge rectifier: 87 C
The rectifier is really hot. But when it still works after more then 40 years (produced in 1981), I can't say it's not fit for the job...


The multimeter is a real gem. It's accuracy is much better than 0,02% (DCV) and it has great looks! Yes, tastes differ... 

[slbender]

Because of the comments, …. The unit was powered on for 45 minutes (235V / 50 Hz):

Transformer coils: 43 C
And of course, the internal of the coils shall be warmer.

Iron part of the transformer: 35 C
So just 2 C higher after 15 minutes more, compared with the first measurement.

Bridge rectifier: 87 C
The rectifier is really hot. But when it still works after more then 40 years (produced in 1981), I can't say it's not fit for the job...


The multimeter is a real gem. It's accuracy is much better than 0,02% (DCV) and it has great looks! Yes, tastes differ... 

Not wishing to steal this thread, but just to make an observation and comment, I have a different pair of units, small sweep / function generators, each rated at consuming 12 Watts and having 1/8 amp slo-blow fuse in the AC line.  So these are sort of comparable to a DMM.  They basically look the same except for the color of the cover.  The older unit model 1200, the later set is 1200A, I have not compared the insides, but suspect they are very close except for some minor improvements, like changing a pair of output transistors or an IC to a later, faster, or more reliable parts.  The older unit blows the primary fuse, and even with twice the  original fuse value, it fails to power up.

At first I suspected a shorted power transformer primary, but both units read between 42.x to 43.x ohms, hot to neutral at the NEMA socket, so there must be an internal short, most likely: a bridge rectifier, or a ‘lytic or tantalum capacitor that has shorted. My point here being - that stressing power supply components is likely to cause a short or failure sooner or later.  In addition, even in a $50,000 massively over built unit, I have encountered burned power supply connectors and failures, so those really hot parts need to be addressed or your vintage “real gem” of a unit will be smoking or dead, sooner, not later.  A word to the wise should be sufficient!

Steven