• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Evaluating Unknown Power Transformers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've been picking up lots of nondescript transformers for amplifier projects, and a fair number of them have unknown ratings. I plan to load the windings to estimate their current rating. What is generally the accepted droop for a tube type transformer winding at max current? I would think it would be around 5% from the zero load voltage, but correct me if this is not the standard practice.
 
5% is a bit too small, that's a value suitable for low voltage low DCR transformers. For high voltage transformer where you can have up to 150ohm DCR for a 300V winding (like a tranny I have here sitting on the desk), you can expect up to 10% voltage drop.

As always, check if the transformer is overheating, smell is a bad sign.
 
You can estimate the rating from the secondary resistance. The primary does come into it too - usually a transformer will be designed for about the same primary and seconday losses. For a start, allow 2 Watts per winding for resistive loss and see what current that is... for larger transformers, maybe 3-4W.

For a heater windings, I use a 6AS7 as a test load (2.5A) and see what the drop is ...
 
It sounds like there should be a compromise here - 5% droop for low voltage/high current windings ( filaments), and 10% for HV windings. Of course, I'm going to back off if the thing gets hot.... What I need now is one of those great, big slider resistors so I can load up the transformer and take my life in my hands.....
What I'll probably actually do is set up a switched resistor matrix like the one I built at work to test power factor correctors. Come to think about it, since the voltage drop is depedent on the winding resistance alone, there's nothing wrong with checking out the transformer at reduced line voltage, making whatever resistor bank I build up more versatile. The only thing wrong with this approach is that you won't be testing the primary at full volt-seconds to detect any skimpiness there, but you can also tell about that by running the transformer unloaded at full input voltage plus a few percent. If the thing gets too hot, then either the design was skimpy, or you've got a problem like a shorted turn. A shorted turn would also show up big time in the unloaded transformer exciting current. Ah, the joys of vintage transformers....
 
Administrator
Joined 2004
Paid Member
Properly designed power transformers for tube applications will generally deliver their rated current at some recognizable nominal voltage - this is particularly true of the filament windings, and in many cases you can also determine the secondary rating as well.

You need to know the nominal voltage rating of the primary which will usually be 110V (usually ancient) , 115V, 117V and on more recent ones 120V.

You will need a variac and an accurate meter or two, and a load box.
As you load the "tut" you should monitor the primary voltage and adjust to maintain it constant at the value chosen.

A fairly easy way to make the determination of primary design voltage would be to identify the 5V rectifier winding. (Usually yellow leads.)

Most rectifiers are either 2A or 3A filaments, although some windings may support two rectifiers. (rare and usually 4A in my experience.)

Through a process of iteration you will find both a primary voltage and filament load resistance that puts you fairly close to 5V. Note this is not absolute because with the other windings unloaded even at the nominal load the voltage might be a few % high, but it should get you close. It should not be less the 5V and should be less than 5% high if you are in the ballpark.

Once you know the primary voltage you can use the same technique to load the 6.3V windings and determine the rated current.

The high voltage secondary is a little trickier, a possibly useful rule of thumb (and like all r-o-t use judgement here) is that the no load to full load regulation may be something like 10%, however in many instances there are relatively standard secondary voltages that may fall out at specific load currents.

(Note that more modern transformers are often designed to 20% no load to full load in order to save both copper and core size, and hence cost - highly likely in inexpensive SS gear imho..)

All of this assumes resistive loads, and should not be considered absolutely definitive. It is not possible to know the insulation class of the particular transformer, and therefore except in the case of the rectifier filament winding I would be tempted to derate the transformer by 10% - 20% just to be safe - this is good practice anyway imo.. :D

Note YMMV, and this is just one possible approach to making the determination.

Needless to say make sure that the possibility of contact with hazardous voltages is minimized and use a fuse in the primary circuit sized appropriately for the transformer being tested.
Caution is always warranted with high voltage.
 
Wrenchone,

You have several pieces of good advice here, it would seem that designs differ from manufacturer to manufacturer.

I have measured known transformers over the years with the very purpose of establishing a rule-of-thumb should I need one. In my experience the average copper loss per secondary seems to have been about 3 - 4%. (That would mean an efficiency of about 85% taking primary copper resistance as well as iron loss into account.) The h.v. winding is then easiest to estimate: You calculate 4% of the open voltage and divide that by the winding resistance to give current, as Tom Bavis indicated. This method is not useful to calculate 5V and 6V currents because of the low winding resistance. In that respect Kevinkr's method is probably best - or just guess what type of circuit the estimated h.v. current would probably support.

There is also a basic equation for estimating the wattage from the core section area, viz. A = [sq.root(V.A)]/5.58 where V.A is the voltamp (watt) and A the core section in sq. inches. I have however found this to be generous.

In the end it would also depend on your climate! In India you would probably need to run transformers cooler than in Canada.

Regards.
 
The rub in India is that the mains voltages go all over the map, with really scary surges. It's a challenge for both linear and switching power supplies. Motorized automatic tap changers are common household items there to help keep the voltage within reasonable bounds.
Australia used to have a reputation for high mains voltages, though that may have been in areas at the head end of a long transmission line.
 
UPS supplies cost lots of money for something big enough to power a household. They may even be wiped out by the same sustained 300V surges they are trying to protect against if they are not designed to handle Indian mains. The motorized tap-changer is a cheap, bullet-proof Neanderthal-style approach that can stand up to sustained abuse.
 
Hi:
Here is the color coding for power transformers from the ARRL Handbook 1965 edition.
Primary leads Black
if tapped:
common Black
tap Black and yellow stripe
finish Black and red stripe

High voltage plate winding Red
Center tap Red and yellow striped

Rectifier filament winding Yellow
Center tap Yellow and blue striped

Filament # 1 Green
C T Green and yellow striped

Fil # 2 Brown
C T Brown and yellow striped

Fil # 3 Slate
C T Slate and yellow striped

Hope this helps
Ed
 
You will need a variac and an accurate meter or two, and a load box.
As you load the "tut" you should monitor the primary voltage and adjust to maintain it constant at the value chosen.


Sorry to resurrect such an old thread but subject matter is fully relevant. Can someone offer a practical design for an appropriate load box for testing HV current ratings. Perhaps a link or sharing of a tried and true DIY box ?


I have numerous unmarked PTs to be tested, all potted type. I will add a few bits to the conversation based on personal observation since most of the discussions have assumed conventional open frame PTs with visible wire leads. With potted types there are only pins. It is very common for relative current ratings to be distinguished by pin size. For example, a 6A filament set may have noticeably larger pins than a 1.5A set. However, there are certainly exceptions. I have Raytheon and Merit PTs with multiple 6.3V sets of differing capacities that all use the same pin size - in this case all the pin sizes are of the higher current handling size. Just looking at them and trying to guess based on pin size would be somewhat misleading since each transformer has at least one 6.3V set rated at only 1A. I have tried to discern what the standard was for correlating pin size to what expected wire ga would have been used on a flying lead type but haven't come up with anything reliable. Given the huge variation across all iron types (low level mic inputs with giant pins for example) I've often wondered if the pin size was used in some cases only as a visual reference.

I have successfully repurposed an old 10 tube organ chassis as a filament load tester. Stock 6V6 pair will take 6L6 & EL34 for a wider range. Pretty handy. Now I need a good simple design for loading the HV. Hopefully someone can help. Lots of good advice in this thread so far.


Thanks,

Alan
 
HV winding will have more sag than you might think since a cap-input rectifier has a high peak-to-rms ratio. But that's what PSU Designer is for... heating depends almost entirely on the RMS current, but voltage drop will be proportional to peak current.

Core area will determine the maximum VA, winding resistance will give a good idea of current capacity.
 
heating depends almost entirely on the RMS current,

Almost, but in some cases the peaks associated with a cap input filter push the core into saturation. When this happens things get HOT quick.

The Triad N-68X isolation transformer is rated for 50VA. With a light bulb load it is warm but not too hot to touch at 50 VA. The same transformer feeding a full wave voltage doubler is smoking hot at 35VA. I had to use a 100VA transformer for that amp design.

You can get a rough estimate of this effect with a Kill-A-Watt power monitor. They can often be purchased from Newegg when they have a sale for about $20. A pure resistive load will show an equal number of VA and Watts. A non PFC computer power supply may consume 1000 VA to produce 600 Watts of DC. The DC supply in a tube amp can be almost as bad.

The guitar amp design I did used series string tubes running off of the 150 VDC from a bridge on the isolation transformer while the B+ was generated with an 8 diode full wave voltage doubler. This is of course worst case since I was using 35 VA to make about 20 Watts of DC at idle. THe 50VA transformer would have fried if I let it run more than 10 minutes.

I still have a fair collection of mystery iron. I have gone through them and marked the measured voltages on a sticker on each transformer. If I have a possible fit, I just try it. I measure the voltage drop from no load to full load on each winding. Look for similar percentages on each winding. A higher drop than all the rest indicates an overloaded winding. Let the transformer run for at least ah hour and see if it's too hot. If you can keep your hand on a vintage transformer, it's too hot. Modern Hammonds can and do get hot. I have a Hammond made Allied in my SSE that gets way to hot to touch, and smells funny after all day use. It is now in it's 6th year of use.

The older transformers have more iron and copper, and thus operate a safe distance away from core saturation. They also have a lower tolerance for heat since high temperature enamel and plastics were not used.

Many modern transformers have been cost reduced to the edge of saturation. Todays high line voltage with high distortion (I measure 10 to 15% with obvious flat topping in the early evening) can push the transformer into saturation when operated within it's ratings.
 
The method in the article in the first post here gets you pretty close...

diytube.com :: View topic - The Dumpster Transformer Topic


Excellent! I've been looking for those curve charts forever. I think I've seen maybe one for a specific commercial transformer but never the compiled one like this. Great find - Thanks!




I tried the method that Johan had outlined earlier (quote below) but am getting results that are all over the place. I tried this method on several mystery PTs and several ones of known ratings. Sometimes the numbers looked correct, other times they seemed way, way off. Not sure how to account for the wild variations.


In my experience the average copper loss per secondary seems to have been about 3 - 4%. (That would mean an efficiency of about 85% taking primary copper resistance as well as iron loss into account.) The h.v. winding is then easiest to estimate: You calculate 4% of the open voltage and divide that by the winding resistance to give current, as Tom Bavis indicated.
 
I recall this from the rusty recess of the gray matter. The small VA transformers are labeled by voltage drop or regulation. I agree that the 5% number you put out is a little low, real world perhaps 20% between no load and full VA with a resistance load (no rectifiers). The larger VA transformers have less surface area for cooling and are limited by temperature rise. I like to use my Fluke IR 62 Mini thermometer and keep the transformer below 140F. In the old days with working man hands 140 would not burn the hand or the insulation.
DT
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.