Signal Amps for output transformer testing

[rackmonkey]

Certain tests used to characterize an output transformer can require more voltage or more power than the typical modern function generator can provide. In his post on determining primary inductance (https://groupdiy.com/index.php?topic=73039.0), CJ mentions increasing voltage level until L stops increasing with it to find the saturation voltage at a particular frequency, for example. A second example: transformer frequency response may look different when the core is fully excited vs when it isn't. So the ability to provide test signals of sufficient voltage or power is key.

Few inexpensive, modern signal/function generators of the type many folks use today are capable of providing more than 7 - 10 volts RMS. Many put out much less. Power output is usually below 125mW (~21dBm). Even most older oscillators output a maximum of 20dBm or less into 600 ohms. For testing output transformers with power ratings  of +24dBm and up, this won't cut it.

A signal amp/function generator amp is useful for these situations. If it's not practical/possible to test using the amplifier the transformer will be used in, or if you're just testing a transformer to see what kind of project it might be useful in, you'll need an amp like this. It should not only deliver the voltage/power you need into the load the primary presents (plus whatever series resistance you add) it should have the frequency response needed (at least down to 10Hz IMO, preferably lower) and not add significant distortion to the signal. A DC amplifier will allow you to inject signals well below the audible band, which is useful to accurately characterize the -3dB low freq point at whatever frequency it might be at. But you may or may not need that capability in your particular scenario.

It can be hard to find a good amp for this job. I bought a cheap chinese signal amp and a couple of old DC amps off ebay, but none of them could provide enough clean output to satisfy the requirements. Thought I'd share what I ended up with.

The two solutions that worked well for me were both dirt cheap.  One is a DIY solution, the other purchased and modified.

My old buddy Pete Millett sells a power op amp board on eBay that was designed for the OPA452/OPA453 power op amp. He sells a pair for $20. You can also use OPA547 or OPA548 chips if want more power. I used the 453 and it works great for pro audio projects. If you need to test transformers for power amp projects, you'll definitely need to upgrade to a more beefy chip.

Link to the boards:
https://www.ebay.com/itm/DIY-PCB-2x-Power-Opamp-Driver-PCB-OPA452-OPA547/150913782253?hash=item23232991ed:g:OmUAAOxyIPNTcmmE

Description on Pete's website:
http://pmillett.com/pwrop.htm

Super easy to build. Took half an hour to stuff the PCB. You need to supply the power. Pete sells a cheap supply PCB with it, but you'll have to design it yourself using PSUD or similar to determine part values. I ended up designing a +/-36 volt supply with it, and it worked for both this amp and the one below.

Second amp is one I came across on the web after stumbling onto an article about finding/building a signal amp. I ended up buying one from Alibaba like the author did, but later they started becoming available on eBay. I think I paid $35 or so.  It's larger than the one above and is based on the OPA541 power amp chip, which is more powerful than the OPA453 (also generates a lot more heat, as you'll see when you read the original article that turned me on to these).

I made the modifications the author of the article below made and I've been happy with the result. It involves swapping out some SMD resistors and removing a cap, IIRC, but I didn't have a problem with it. You could also just use the amp as is if you don't need DC amplification or don't want to set the gain at 10.

https://www.dmcinfo.com/latest-thinking/blog/id/9462/low-cost-function-generator-amplifier-diy

You can find those all over eBay now, and even on Amazon. Be sure they have the chip you want in them, as I see them being sold now with any of several different TI power amp chips.

Simple projects that yielded good solutions to the problem.

 

[Tubetec]

They do seem like good value for money , your typical ,even Chinese instrumentation grade power amp module costs hundreds , a decent enclosure and a psu is all thats required to finish that module into something nice .
Looks like the OPA541 is  capable of driving down to  speaker type impedences at least for short term test signals  which is  a useful feature to have .

Theres an alternate version with the OPA541 as well ,
maybe slightly better lay out .

https://www.ebay.com/itm/OPA541-module-HiFi-audio-power-amplifier-board-High-current-5A/273595424664?hash=item3fb38edf98:geAAAOSwFx5cB4OH


Im only barely scratching the surface of what programs such as REW can do for transformer measurement ,watching how the proportions of the various harmonics change as level/distortion is increased seems to reveal everything about how  the core behaves .

 

[emrr]

Nice, thanks.

 

[rackmonkey]

You could also use these OPA4xx/5xx chipamps to make a nice quality headphone amp. Another set of options beyond the usual suspects like the LM48xx/49xx ICs.

Read PMillett’s findings/musings in the page I linked to at the top.

 

[Tubetec]

I couldn't resist buying a couple of HP 200CD gens I saw on offer just for the high voltage drive into >600 ohms ,and the chalenge of bringing them back to life.

Would it be possible to rig aforementioned two stage op amp based power board  with a sharp hi cut  filter,  just like we see in the ultra low distortion type signal sources , then your medochre sound card output could  become an automated  software controlled low distortion signal source ,capable of driving almost any load .
Spectral display of the harmonics tells its own story , once you have the main channel and loop back cal'd out , its like having 'Einstein' in a box doing the mathematical heavy lifting for you 

 

[alexc]

I have some 'amp modules' based on that amazing Texas Instruments LME49811 'power amp driver chip', a TIP41 bias transistor  and a pair  TIP147/142  complimentary silicon darlington power transistors for as much power as you can sink.

This small and relatively cheap module (10cm x 5cm 25usd) has edge a heatsink for the TI chip and the transistors are edge mounted for easy heatsinking, as massive or fan assisted as you like.

For the real loonies, you can 'parallel TIP devices' to make whatever power you have 'in mind'.

On my bench, I like to make something like 70V ac signal into various 'balancing transformer' loads ...  the super low THD+N from the 'amp module' combined with the 'step up' from a suitable audio transformer gives me a great, super-duper hifi large amplitude ac signal with which to test 'tube output stages' 

...  it comes out more hifi than any other method I know for these 'neutral sounding'  70 .. 200V ac signal swings  one uses to evaluate big output stages like p-p 6L6 or 6550  and even to the beyond  ..  a pre-requisite for some of the more esoteric single-ended amp builds.

Once you know what the output tubes are capable of, you can really hone your 'driver' sections. 

..

So yes - serious fidelity signal amps at reasonable powers are a great thing to have