Age-old question: reverse, perverse engineering of surplus power xfmrs.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
OK, not a totally unknown 'found' transformer, but an incompletely-specified transformer I have boxes of (NOS).

I know what the primary and secondary voltages are from a schematic of a product that used this specific transformer, but there are no current specs.

So I took a notepad and collected some measurements. (Weight, dimensions, partial destructive disassembly- sacrificed one to learn more).

Heathkit 54-835, one of two transformers supplying power in the GR-2000 color television. Made by Midwest Transformer (EIA code 549) in Gray's Lake, IL, 1982 (Co. moved to Mexico in mid-1990's, out of business by late 1990's, per inquiry to another competing manufacturer).

~8#/3.6 kg, EI-137 laminations, 0.019", probably M-19 26 gauge steel, based on Tempel catalog. 4.13" x 1.84" x 3.45" iron. 1-3/8" tongue width, 1.84" stack, 2.53 in^2 'stack'. Removed end bells (0.304# each and removed flying leads during process of cutting insulation to measure magnet wire). Impregnated with some horrible black stuff apparently called Perma-Fil, may be an epoxy. There is varnish under the outer fishpaper wrap. Unknown quantity of black potting chipped away with screwdriver. Resulting open transformer weight 7.14# (3.25 kg).

Primary: 120 V 60 Hz nominal, with HI and NORM taps. These three leads go to magnet wire with diameter of 0.034", which I will call AWG 20, circular mils = 1024.

Three secondaries (one is useless):
1) (Useless one) 1.3 V which combines with 2nd transformer 5 V winding to allow CRT to be preheated at 5 V & operated at 6.3 (instant on feature, I guess). current rating unknown, but it was for a 25" color CRT; I'll guess 0.6 or 1.2 A (but wire gauge seems small: 0.011-0.012" (conservatively call it AWG 30 instead of 29 or 29-1/2; 112 c.m.) < 2 VA; just ignore it.

2) 59 VCT (29.5-0-29.5), all leads 0.020", call it AWG 25, 320 c.m. Produces 77 VDC with FWB.

3) 200 V with tap creating 125 and 75 v windings (0-75-200). Used for multiple DC voltages using FWB and HW rectification. 0-75 winding 0.015-0.016" wire, AWG 27, 202 c.m. 125-200 winding 0.032" wire, AWG 21, 812 c.m.
--------------------------------------------------------
A recent Electronic Design News article on transformer design suggested using 500 c.m./A to estimate wire circular mil area or current rating, working backward, and using a lower or higher number for different scale transformers. 1500 c.m./A proposed by some is fine for DIY people with particular goals, but I am looking for practical values the transformer manufacturer would have used.

I took 500, 700 and 400 c.m./A targets and the measured c.m values to estimate mA 'ratings', for each winding.

The TV had a 2.92 A circuit breaker. Not knowing the load distribution shared by two dissimilar power transformers, I can only estimate half the VA for each of them, and allow 50% operating current relative to the breaker nominal rating. 120*1.46 =175 VA on the primary. Rule of thumb estimate of 90% efficiency=157 VA secondary. 40 VA/kg rule of thumb ~130 VA.

This seems kind of lame for an 8# transformer, hopefully closer to overly conservative than paranoid superstition.

I expected to find winding resistances in the ten's of ohms range, but the highest was only 3.5 ohms, suggesting decent current rating.

Once I 'accept' a total VA rating estimate, I'll choose an estimated current rating for each winding based on c.m./A and reduce it by roughly 1.8 to allow for current crest factor resulting from FWB and capacitor input filtering (smallest capacitor was 100 uF, and other supplies used 200 and 300 uF) to estimate VA for each winding, watching the total calculated VA resulting for the two useful secondaries to not exceed some the total secondary VA estimate.

Maybe put the 200 and 59 vac windings in series and use the smallest winding wire gauge as the worst case. At 700 c.m./A, that would allow 0.286 A. 0.286*259 is only 74 VA. At 400 c.m./A, 0.50A, 259*0.5 = 130 VA.

I'm whittling these down to the level of conservatively serving as doorstops. Plan B is to just calculate some resistive loads for the DC rectified/filtered case (~365 VDC unloaded) and let the transformer cook for a some hours (24?), and measure the increase in the highest winding resistance to estimate temperature rise internally.

Gotta be good for some kind of B+ supply. Scrap dealer will only pay for the weight in steel because they have to do the work to get the copper out.

Any interpretive opinions are appreciated. (Hey, I could have worse hobbies).

Thanks
 
Last edited:
One more thing worth mentioning: all the information I gave pertains to 50Hz Xformers.
In a 60Hz system, the powers estimated could be increased by ~20%

Yet another thing: the A value should be the net area, not the geometrically measured, physical area: because of the insulation of the laminations, the deformation due to the stamping process, etc. the stacking is imperfect, meaning it should be reduced by a factor (called "foisonnement" in French, for which I am not going to venture into providing a translation) generally taken as 10%, meaning the power should be reduced in the same proportion
 
Last edited:
Thanks for add'l input.

I borrowed a Kelvin-lead-equipped milliohmmeter today and took winding resistances for two different p/n transformers made by the same supplier. Both from different model color TV's, but different combinations of windings.

The second one I selected because I have the specs for that one and perhaps being from same mfr., made for same customer for similar application may add some confidence to the estimation.

Have to re-measure unloaded voltages again...can't find that notepad.
 
Last edited:
2nd transformer produced secondary/primary V2^/R 'constants' ratio of about 0.66, lower than I expected but at least I expected it to be less than unity. This transformer is intended to be used with a large fraction of its load using a voltage doubler and a significant but smaller fraction for tube/valve filaments. I expected some scaling phenomena with the rectifier part because of current crest factor (I think that's the right term).

The first transformer produced something very strange so I suspect measurement or data logging errors on my part. All the unexpected results were on one winding, so I will repeat those 4-wire resistance measurements. I thought they were suspiciously low when I wrote them down (compared to 4-1/2 digit 2-wire DMM measurements with lead resistance zeroed out taken prior to reading the recommendation for Kelvin measurements). I expected discrepancies on the lowest resistance windings and was surprised the highest resistance windings were where I had very significant discrepancy.

It was definitely reassuring to see stable and repeatable measurements to 1 or 10 milliohm resolution so I'm waiting to re-measure the suspicious winding Friday. The only oddity about that winding (0-125-200 vac) is there are two magnet wire gauges so obviously the load is surely different on the two windings.

One other difference on this transformer is that it has a normal and high line tap for the (single) primary. I applied 115.1 vac to the normal tap and 'neutral' wires and measured the full winding (essentially as an autotransformer if looking at the primary only), and the full winding (hi to neutral or start to finish) was 122.4 vac.

The reason I am spamming the forum with this is because SOMEBODY out there might find it educational, and if not, a cure for insomnia...

Both meters used belong to a metrology lab a friend works at and have current valid periodic calibration certification (so it's probably ME).
 
The drastic discrepancy was my error...the milliohmmeter had a 2nd range switch for the lowest range, and when I repeated the measurements with a different meter (same model) I realized my numbers were of by a factor of ten, which would screw up the V^2/R results.

The second set of numbers made sense, and I also tried an LCR bridge (analyzer) with a DCR mode and also Kelvin clips got remarkably close measurements with both instruments (as I would wish for).

2nd p/n transformer I have specs for so I did not tear one apart for magnet wire gauges.

That one had a primary V^2/R 'factor' of about 18600.

The V^2/R sum of the three secondaries (one for B+ with voltage doubler, and two for unrectified filament heater use) was about 12420...approximately 67% of the primary's.

Interestingly, the smaller transformer with a tapped (high and normal 115 V) primary and three secondaries (only one was a very small filament winding) had a secondary V^2/R 'factor' of about 12400 also, but the primary calculation was less, about 7700. That's the disturbing thing...12400/7700 is 1.6, not less that 1.0.

I typed up a spreadsheet for each transformer and saved as jpegs but haven't stored them anywhere yet.
 
The V^2/R sum of the three secondaries (one for B+ with voltage doubler, and two for unrectified filament heater use) was about 12420...approximately 67% of the primary's.
It could be caused by a sloppy design (or the need to use certain wire gauges), or by very poor quality iron requiring a disproportionate magnetising current

Interestingly, the smaller transformer with a tapped (high and normal 115 V) primary and three secondaries (only one was a very small filament winding) had a secondary V^2/R 'factor' of about 12400 also, but the primary calculation was less, about 7700. That's the disturbing thing...12400/7700 is 1.6, not less that 1.0.
The tap is too small to have a significant influence on the global result: things would be different for older multivoltage primaries: 110, 120, 140, 190, 220, 240V.
An explanation could be that not all secondaries are meant to be used simultaneously
 
Thank you.

The smaller one does produce multiple voltages from one winding (bridge and reversed HWR) in addition to powering a degaussing coil in parallel with a thermistor in series with the winding before it reaches the bridge.

I spent some time today comparing some filament winding ac VA to the windings that have a bridge or doubler to 'select' a derating factor. I use a VA reduction factor of 1.8 for bridge rectifier, and decided I should double that for doubler. I have not seen that in writing anywhere but it seems to 'adjust' the V^2/R to VA ratio to be in the range of the other windings' ratios.

Then I compare the c.m. value for each winding. I'll be choosing 'my' current rating and then find some load resistors next and see how warm it gets.
 
These also have
some of the the 'old-school' characteristics you mentioned...

In some cases a tapped primary, overlaid windings rather than split bobbin, and a maximum temperature rise with some 'test circuit'. The max. core losses & magnetizing current are also specified for the one I found data for.

Thanks for your help.
Hopefully, I'll choose my factors safely & conservatively.
 
The smaller one does produce multiple voltages from one winding (bridge and reversed HWR) in addition to powering a degaussing coil in parallel with a thermistor in series with the winding before it reaches the bridge.
For the degaussing winding, it does not "exist" in the original planning of the transformer, because of the thermal time-constant.
Of course, if you decide to use it on a permanent basis, it has to be part of your new global calculation.

I use a VA reduction factor of 1.8 for bridge rectifier, and decided I should double that for doubler. I have not seen that in writing anywhere but it seems to 'adjust' the V^2/R to VA ratio to be in the range of the other windings' ratios.
Contrary to a common belief, the PF of a doubler is not very different from a fullwave rectifier.
What I mean is that a winding capable of providing 12V/1A DC after rect+filt will deliver practically 24V/0.5 DC with any type of doubler.
I have made a thread on the subject several years ago


These also have
some of the the 'old-school' characteristics you mentioned...

In some cases a tapped primary, overlaid windings rather than split bobbin, and a maximum temperature rise with some 'test circuit'. The max. core losses & magnetizing current are also specified for the one I found data for.

Thanks for your help.
Hopefully, I'll choose my factors safely & conservatively.
With overlaid windings, the primary is generally the innermost winding, meaning it will mechanically have a higher power coefficient because of the shorter turn length, yet it will be the most susceptible to heating, so all of that has been said has to be taken with an open mind, a pinch of salt and a good portion of context.
 
I found 24-year-old notes dating to when I actually had data for this no-longer-mystery transformer. Now I can compare specs to estimates.

primary 60 Hz 0-120(normal line)-128(high line)

sec. #1 derives two DC voltages
1) 59vct FWB across full winding(CT floating)+600 uF 77VDC@0.18ADC 0.020" AWG25 320 c.m.
2) ct to 500 uF produces 37.5 VDC @ 0.28A (appears to be HWR by one leg of the FWB)

sec#2 was drawn in notebook as two windings internally, and I did find two wire gauges. Schematic shows 0-125-200vac.
0-125 is 0.032", AWG21, 812 c.m. applied to FWB+300 uF: 158VDC@0.77ADC
'upper' 75 (125-200) is 0.015", AWG27, 202 c.m., HWR+100 uF adds ~100VDC in series with 158VDC: 268VDC/0.025ADC

sec#3 is strange. 1.3vac/0.9a ac, was in series with 5vac/0.9aac winding on 2nd transformer (I don't have) for CRT preheating/instant-on feature. Useless on this xfmr by itself, and wire is only 0.011", AWG30 100 c.m. (my diameter measurement was reference for c.m. and AWG 'labels'). Seems undersized by any c.m./A rule of thumb.

Looking at DC current ratings and contemplating c.m./A, I'm reminded that it is probably properly applicable to transformer AC VA rating and is skewed by rectification and filter capacitor size...the only winding I would want to 'squeeze' more out of is the 75 vac AWG27 winding in series with the 125 vac AWG21 one. If I used the 0-125-200 'series' winding with ONE FWB across the full dual winding, maybe 268 VDC upped to 200 mA. Or use the 158 VDC/0.77ADC winding with a doubler at half the current (or derate that a little if uptight), or move the ground point to have 75-0-125 vac, with a doubler on the 125 and FWB on the low-current 75vac winding & use for filament supply elevation for totem-pole circuit upper tubes, or negative HWR for a bias supply.

Murray
 
Or just put all windings in series:
125+75+59=259vac. The 0.18ADC rating of the 59 v winding is close enough to my target of 0.2ADC for the 125+75, limited by the 75v AWG27 winding.

Been saving a couple cartons of these since 1995, and they get heavier every year!

I have separate 6.3 V filament transformers.
 
Hi Murray,

I just have one little tip to add -- and it sounds like you came around to it already --

On 'compound' windings (where wire size changes at a tap) I always assign my '0' toward the heavier-gauge end.

Also, sure feel ya' on the 'heavier every year'. Thinking of how much harder it got between our last two moves, the next one gives me a real chill. And wives notice those cartons, too -- and how heavy they are!;)

Regards, Rick
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.