Output transformer education wanted...

[mschmitzusa]

I have read a lot about output transformers lately and learned a lot. However there are a few things I have difficulties with. Here are some questions coming up. I have read a lot about paralleling output transformers to get more power and also the ability to get odd impedances etc. Here is my story: I am in the process of building some sweep tube amp and thought of like a Bogen DB130 clone and maybe even double up on output tubes for more Watts. The problem is that universal outputs available are made either for table radios or guitar amplifiers. I want Hifi, but the regular production output transformers are made for the commonly used output tubes only and do not have the impedances I need. A custom winding job I can't afford, just like I can't afford to pay hundreds of dollars for a quad of EL34 or KT77 or KT88 tubes after the sanctions on Russia and the resulting quadrupling of output tube prices, where as sweep tubes cost 3 to 5USD NOS a piece. I heard of people paralleling lower grade (guitar or console) output transformers for example and only loading them to about 50% of the transformer VA rating and claiming to improve frequency response due to increased iron and copper without changing in and output impedance. I spoke to Edcor and Hammond, but since they are in the selling transformer business, they were not very helpful and tried to push me into the very expensive custom winding job realm. I found an Edcor XPP15-7.6K with 6Ohm output and 70Hz to 18kHz at 1dB output, which has a 6Ohm output that I could load with 8Ohm instead to get my desired primary impedance of 11400kOhm. Unfortunately, that transformer is only 15Watt and I need very close to 40Watt. I asked Edcor and did not get straight or enlightening out of the box answers. Now I am asking the Audiokarma community, if I parallel three of them to get 45Watts output, will the increased iron core made up of 3 cores increase my low and high frequency into the 20 to 20kHz at 1dB or close by area? Six of those transformers can buy for what Edcor charges as setup fee alone to make a custom winding job, you still have to pay on top for. Kindly advise and give me a good lesson on that subject. Thanks!

 

[Lavane]

I don't think I've ever seen an amp with paralleled output transformers.
Why not use a single output transformer made to the output needed and just parallel tubes to get there?

 

[maxhifi]

If you increase the core area, you will improve the low frequency performance, but only if you don't also increase the power output of the amplifier. Nothing will improve high frequency performance except for a better transformer with more interleaving of the windings.

So say you have one 15W transformer which can handle 70Hz - 18kHz. Two in parallel will get you 15W at 35Hz - 18kHz, or 30W at 70Hz - 18kHz.

I suggest looking for a used amplifier if cost is an issue, since they can often be obtained for less than the value of the parts.

 

[mschmitzusa]

Custom winding: 300USD setup fee, minimum order of 5 transformers at 200USD equals 1300USD to get the proper transformers made. If doubling up on tubes to get down the primary impedance you need a 5700k output, not commercially made either. It would be either 5000k or 6600k and that requires making changes in the circuit to match it up properly. I never designed an amp at that level. The Bogen DB130 is a good sweep tube amp with excellent data and reviews, but the people selling it on the Bay won't even warranty the transformers, but want 200+USD for a beat up amplifier carcass plus 150USD shipping and then I have to freight forward that for another 150UD to Jamaica where I then have to pay 70% duty on ALL the cost inclusive shipping as added value.

 

[Tom Bavis]

Paralleling transformers doesn't change the impedance since the voltage ratio determines that. And doesn't help the low end since that's determined by the saturation point (voltage / frequency). Primary inductance is cut in half, but not a problem since most transformers are limited by saturation, not inductance. High frequency response could be better since leakage inductance is cut in half. It will increase power handling capability and reduce losses, since current is split between transformers.

If your transformer is higher than the impedance shown on the data sheet, decreasing the screen voltage can optimize the operating point, and vice-versa, within limits. But a 10% impedance change will make only a few percent change in max output. If you're a volume manufacturer, you'll optimize the circuit since competitors can do better at the same price if you don't. If you build your own, the economics can be different - if I can do 95% of the performance at half the price, I know what I'll choose!.

 

[gadget73]

I don't think I've ever seen an amp with paralleled output transformers.

closest I can think of is something like the Bogen MO-200A, which is effectively two 100 watt power amps on one chassis with the input and outputs in parallel to make one 200 watt amp. Same thing that happens with a Dyna ST-70 if you flip the mono switch and jumper the output terminals together like the instructions say you can do.

I suppose in theory you could build something and parallel the plate and speaker sides of the transformer but somehow I suspect that would turn into a real mess to stabilize.

anyway, just curious where you got 11.4k for the sweep tube load from.

 

[trainbuftony]

Custom winding: 300USD setup fee, minimum order of 5 transformers at 200USD equals 1300USD to get the proper transformers made. If doubling up on tubes to get down the primary impedance you need a 5700k output, not commercially made either. It would be either 5000k or 6600k and that requires making changes in the circuit to match it up properly. I never designed an amp at that level. The Bogen DB130 is a good sweep tube amp with excellent data and reviews, but the people selling it on the Bay won't even warranty the transformers, but want 200+USD for a beat up amplifier carcass plus 150USD shipping and then I have to freight forward that for another 150UD to Jamaica where I then have to pay 70% duty on ALL the cost inclusive shipping as added value.

Shipping to Jamaica is no joke, they steal packages all the time in addition to the 100% duty. If I were in your shoes, I'd be looking for a transformer winder locally who can make what you need. Given the photos of enormous Bass guitar amplifiers and Sound system amplifiers I've seen come out of Jamaica, there's got to be someone down there keeping that tradition alive. I couldn't begin to speculate how you're going to find them...
In terms of the "education" I'm afraid I'm of little help, I've just scratched the surface of the topic. I can tell you for certain that most of the folks that held the knowledge, are dead...
Good luck.

 

[maxhifi]

Custom winding: 300USD setup fee, minimum order of 5 transformers at 200USD equals 1300USD to get the proper transformers made. If doubling up on tubes to get down the primary impedance you need a 5700k output, not commercially made either. It would be either 5000k or 6600k and that requires making changes in the circuit to match it up properly. I never designed an amp at that level. The Bogen DB130 is a good sweep tube amp with excellent data and reviews, but the people selling it on the Bay won't even warranty the transformers, but want 200+USD for a beat up amplifier carcass plus 150USD shipping and then I have to freight forward that for another 150UD to Jamaica where I then have to pay 70% duty on ALL the cost inclusive shipping as added value.

I think you're approaching this design backwards. First of all, the impedance doesn't need to be exact, because speaker impedance isn't exact. For example, if a tube requires a 5000 ohm load, then a transformer 4000 - 6000 is all fine.

Secondly, it makes no sense to get custom parts if the aim is to save money.

What sort of amplifier are you trying to build anyway? Sweep tubes and ~ 50W output? I suggest to check diyaudio.com, there's some guys there using inexpensive toroidal power transformers as output transformers with Russian sweep tubes and getting great results. It might be up your alley.

 

[Lavane]

Pete Milmette sells boards for his Engineer's Amp that uses sweep tubes for output and RF signal tubes for drivers. They can be built 2 channel, or mono blocks. They're only 20 or so wpc though. These amps use off the shelf transformers.

 

[zackthedog]

Just to be clear, you're looking for 11.4K primary at 40 watts? If so, Heyboer's Peerless S-265-Q (10K primary@40 watts) copies would be close enough, I think. Or perhaps I misunderstood what you're looking for...

 

[Mr Boochie]

I have read a lot about output transformers lately and learned a lot. However there are a few things I have difficulties with. Here are some questions coming up. I have read a lot about paralleling output transformers to get more power and also the ability to get odd impedances etc. Here is my story: I am in the process of building some sweep tube amp and thought of like a Bogen DB130 clone and maybe even double up on output tubes for more Watts. The problem is that universal outputs available are made either for table radios or guitar amplifiers. I want Hifi, but the regular production output transformers are made for the commonly used output tubes only and do not have the impedances I need. A custom winding job I can't afford, just like I can't afford to pay hundreds of dollars for a quad of EL34 or KT77 or KT88 tubes after the sanctions on Russia and the resulting quadrupling of output tube prices, where as sweep tubes cost 3 to 5USD NOS a piece. I heard of people paralleling lower grade (guitar or console) output transformers for example and only loading them to about 50% of the transformer VA rating and claiming to improve frequency response due to increased iron and copper without changing in and output impedance. I spoke to Edcor and Hammond, but since they are in the selling transformer business, they were not very helpful and tried to push me into the very expensive custom winding job realm. I found an Edcor XPP15-7.6K with 6Ohm output and 70Hz to 18kHz at 1dB output, which has a 6Ohm output that I could load with 8Ohm instead to get my desired primary impedance of 11400kOhm. Unfortunately, that transformer is only 15Watt and I need very close to 40Watt. I asked Edcor and did not get straight or enlightening out of the box answers. Now I am asking the Audiokarma community, if I parallel three of them to get 45Watts output, will the increased iron core made up of 3 cores increase my low and high frequency into the 20 to 20kHz at 1dB or close by area? Six of those transformers can buy for what Edcor charges as setup fee alone to make a custom winding job, you still have to pay on top for. Kindly advise and give me a good lesson on that subject. Thanks!

If you want more clean power, it is best to go with the bigger output tubes and corresponding output transformers to match them. Don't forget you will need a bigger power supply with more reserve capacity, too. Bass and clean headroom need reserve power.

Laws of Physics. Sorry.

 

[gadget73]

So here is a wierd theoretical ququestion

Lets say we're building a 6bq5 amp, so 8k plate to plate. If we do PPP it wants 4k plate to plate right?

Instead of tying the plates together and running to a single 4k, could you use two 8k transformers, not tie the plates, but tie the speaker side together instead?

Setting aside any stability nightmare crap

 

[Mr Boochie]

I've never seen paralleled output transformers in an amp.

I found this article while Googling.

https://groupdiy.com/threads/two-transformers-as-one-series-parallel.1169/

 

[opnly bafld]

The original Decware Torii (rare blue amp ~12 watts with EL34s) used 2 single ended OPTs in series for push pull.
Not many were built because of the difficulty of getting the 2 OPTs to work together, so the story goes.

 

[TriodeLuvr]

sweep tubes cost 3 to 5USD NOS a piece

Which tube types are those? A pair of 6P3S-E tubes (5881 equivalent) can produce 40W with an off-the-shelf transformer. This tube is still available at good prices from sources in Europe.

Jack

 

[FlaCharlie]

I think you're confused about the OT specs that you need.

You don't mention what tubes you plan on using but you say:

I am in the process of building some sweep tube amp and thought of like a Bogen DB130 clone and maybe even double up on output tubes for more Watts. . . . I found an Edcor XPP15-7.6K with 6Ohm output and 70Hz to 18kHz at 1dB output, which has a 6Ohm output that I could load with 8Ohm instead to get my desired primary impedance of 11400kOhm.

It looks like the Bogen DB130 used 6AV5 output tubes in PP. A couple of threads about DIY builds using these in PP seem to indicate that the OT primary is typically in the range of 5k to 6.6k.

If you plan on "doubling up" that would mean running parallel push pull (PPP). When you run tubes in PPP you run a primary impedance that's half that of a PP build, not double. So you'd be looking for something in the range of 2.5k to perhaps 3.5k.

Hopefully you also understand that doubling the power output does not result in an amp that's "twice as loud". The result of doubling the power output is an increase of 3db which is the equivalent of standing 1m (39") closer to your speakers. Noticeable, but not a massive increase. It takes 10x the power output to make something sound, subjectively, "twice as loud".

In most cases, the result of higher maximum output power is just an increase in headroom. If your normal listening level only requires 1 or 2w then increasing the maximum power from 10w to 20w or 40w only affects the headroom since your normal listening level still only requires 1 or 2w. Nothing wrong with increased headroom but it matters more with music that is less compressed and many modern recordings use quite a bit of compression.

Even if the tubes are cheap, going from a PP to a PPP design will result in increased cost for transformers that can handle the additional current. So if you're on a tight budget you might want to consider how important the increased headroom is to you.

 

[TriodeLuvr]

Even if the tubes are cheap, going from a PP to a PPP design will result in increased cost for transformers that can handle the additional current.

Well, I have to disagree with this. Halving the transformation ratio might require a larger gauge primary wire, but it also means 1/4 the number of turns. Not only does this come at no increase to manufacturing cost, it carries benefits relative to resonance and high frequency response. The issue for PPP really is the cost and availability of tubes. That's especially true if you're like me and only use NOS Western and Russian tubes.

Jack

 

[FlaCharlie]

Well, I have to disagree with this. Halving the transformation ratio might require a larger gauge primary wire, but it also means 1/4 the number of turns. Not only does this come at no increase to manufacturing cost, it carries benefits relative to resonance and high frequency response. The issue for PPP really is the cost and availability of tubes. That's especially true if you're like me and only use NOS Western and Russian tubes.

Jack

I was thinking more in terms of having to use both power and output transformers that are suitable for a circuit that's using 2x the current. From what I've seen higher current specs generally cost more.

Looks like DIY in Jamaica is an expensive proposition if you're importing parts. Since the OP hasn't responded I have to wonder if he is still interested.

 

[GordonW]

So here is a wierd theoretical ququestion

Lets say we're building a 6bq5 amp, so 8k plate to plate. If we do PPP it wants 4k plate to plate right?

Instead of tying the plates together and running to a single 4k, could you use two 8k transformers, not tie the plates, but tie the speaker side together instead?

Setting aside any stability nightmare crap

I did exactly that. Used four of the Leslie 100-12 transformers- 8.5K into 16 ohms each. I tied the two 16 ohm secondaries together to drive a 8 ohm speaker load, and drove each primary from a normal pair of EL84 output tubes. Each pair of output tubes had its own split-load inverter tube, with a common voltage amp per channel driving the two inverters.

Made a VERY STOUT 30 watts per channel. I got lucky, in that there were no significant stability problems- that's NOT guaranteed to happen. But, NOT tying the primaries together, did reduce the interaction of the transformers, and the potential stability issues that could easily cause...

Here's a write-up on that exact amp: https://audiokarma.org/forums/index...amplifier-now-were-playing-with-power.798989/

Regards,
Gordon.

 

[gadget73]

Interesting, i was thinking common inverter but split inverters makes loading less of a concern since the grid resistors are not in parallel and small signal tubes can run fairly high value grid resistors without issue.

So it does work, nice to confirm the theory. I figured it would be more or less like running two amps in parallel for double output just with the common point being a little downstream of the input jacks.