With the inputs shorted I still see the same level of 60 Hz (first attachment). If I remove the driver tubes it is less (second attachment).
Attachments
Hooked up an old 9VDC wall wart to an ebay adjustable regulator and dialed in 6.3V. With this feeding the driver tube heaters the 60Hz noise sits about -110dB so an improvement of 10dB. I'll give it a few more tests and a listen tomorrow.
Ciao Roberto,
Very good question! I would like to know too, and hope George would give his take on it. I already own a pair of the TTG-KT88SE, but wonder if the 1.5k PSE version would be better.
Looking at the specs provides some curious information. The PSE has almost half the primary resistance of the SE, (probably thicker wire for higher current handling in the PSE and fewer turns?) yet the inductance of the PSE is higher (46H vs. 39H). I expected the SE to have higher inductance than the PSE. Does that imply the PSE has more turns than the SE on the primary, and even more turns on the secondary to maintain the design turns-ratio?
Whatever the answer to that question, on the secondary side I assume the two transformers both have a single 8 Ohm coil (interleaved of course) that is tapped for 4 ohm. If so, that means part of the coil is unused if you connect a 8 Ohm load to the 4 Ohm tap. I always thought it is better to engage all (or as much as possible) of the secondary coil. (I thought that was one of Lundahl’s “secret sauces” because their transformers are always connected to use the entire secondary coil.) So, in my opinion, connecting a 8 Ohm speaker to the 4 Ohm tap is not optimal.
So, we have some unanswered questions about how Toroidy constructs these transformers. It might be best to ask Toroidy which transformer will provide the best performance in this application. I believe you have communicated with Toroidy directly, ask them
Good news. Interesting progress. I assume that was with a driver tube installed. What happened to the 60hz without a driver tube, when you had the heater DC connected to the board?
Let us know how you listening and testing go.
Last edited:
To be clear only the driver tubes have the heaters hooked up to DC, output tubes are still hooked to AC. With the driver tubes pulled the results are the same. Attached are the measurements with the driver tubes in hooked to DC and input disconnected.
Attachments
Thanks for the reply. Could you elaborate please.
1. Did Toroidy explain their “PSE is a good idea” recommendation? I would like to understand why.
2. If Toroidy were to build the “optimal transformer” how would it be different in specifications? If there were significant benefits to a custom transformer we could think of a group buy - I expect the 40-50% premium you were quoted is for a “one off”, so I hope a group buy would make it more affordable for a lot of folks.
1. Did Toroidy explain their “PSE is a good idea” recommendation? I would like to understand why.
2. If Toroidy were to build the “optimal transformer” how would it be different in specifications? If there were significant benefits to a custom transformer we could think of a group buy - I expect the 40-50% premium you were quoted is for a “one off”, so I hope a group buy would make it more affordable for a lot of folks.
I will for what I know and what I guess.
1. AFAIK their bigger "iron" is the one they have in the KT88 SE and PSE, and it's in the 3 kg league. Reccommend a standard OPT that costs 100 € will put all major competitor away. Recommend a transformer that costs 140-150€ in the 3 kg league will open the doors to 5 kg league that can be in the same range of prices (around 170 €, but when I was a frequent buyer I had a 15% discount, and this made them back in the game).
So I guess the anwer is mostly commercial on point 1.
2. I would ask them to reduces losses of course, so Rdc of the primary and secondary will be reduced, but I would wait George's feedback on the core, to understand if it is reasonable to approach the group buy, or other options will be preferrable.
1. AFAIK their bigger "iron" is the one they have in the KT88 SE and PSE, and it's in the 3 kg league. Reccommend a standard OPT that costs 100 € will put all major competitor away. Recommend a transformer that costs 140-150€ in the 3 kg league will open the doors to 5 kg league that can be in the same range of prices (around 170 €, but when I was a frequent buyer I had a 15% discount, and this made them back in the game).
So I guess the anwer is mostly commercial on point 1.
2. I would ask them to reduces losses of course, so Rdc of the primary and secondary will be reduced, but I would wait George's feedback on the core, to understand if it is reasonable to approach the group buy, or other options will be preferrable.
Unfortunately I have little to offer here.
Generally an OPT performs best when used under the conditions it was designed for, assuming that it was properly designed and realistic specs were quoted.
I have found exceptions to this though. Over 20 years ago I bought 200 "6600 ohm to 0-4-8-16 ohm, 80 VA, 80 Hz to 5 KHz" output transformers that were designed for guitar amps. On the surface these would suck for HiFi applications, but I have found plenty of places where they work quite well. Most of them were where they were operated as 3300 ohm to 0-2-4-8 ohm OPT's. This is not the usual case though.
I had heard of Toroidy OPT's but never seen one until I bought the pair I have here in post #19:
SE Output xfmrs....who are the current suppliers for not many $$$
My intention was to run them in a sweep tube SE am at 50 WPC or so which requires a fairly low load impedance. This is why I got the PSE version.
They will not make 50 watts without severe saturation at low frequencies. I have found a way to make them work at 25 watts or so (the UNSET), and in this manner they perform far better than anything else that I could get for the same price.
My pair will be going into an UNSET based 20 WPC amp, unless my PSE sweep tube experiment works out, but major workbench repairs have ended all tube amp work for the immediate future.
See posts #2162, 2169, 2176, and 2184 here:
What did you last repair??
At to whether the SE, or the PSE version would work the best in a 3000 ohm application I can't say without testing both side by side in the same amp where they will be used.
I have no information as to how these transformers were made, and I am not a transformer expert. All my DIY OPT experiments years ago were not worth using, so I gave up on that venture.
If Toroidy told you that the PSE would work better than the SE, I would respect their opinion over my "guess."
Thanks George, I remember your posts about those guitar transformers, and I also remember you noticed a notch in the audio range for Hammond SE transformers.
I will ask Thomasz a price for 20 pairs (will we reach that number?) of 3k to 4-6-8 Ohm with UL taps with... which targets we want to reach on the windings?
Lower UL tap than standard 40% (for those who want to switch between pentode and UL)?
Rdc on primary lower than 50 Ohm?
Rdc on secondary lower than 200 mOhm?
Capacitance (please note that this is tricky, as it is high in value, but limited to a small percetage of the winding, so being equivalent to a smaller capacitance)?
Primary inductance?
Other parameters?
I will ask Thomasz a price for 20 pairs (will we reach that number?) of 3k to 4-6-8 Ohm with UL taps with... which targets we want to reach on the windings?
Lower UL tap than standard 40% (for those who want to switch between pentode and UL)?
Rdc on primary lower than 50 Ohm?
Rdc on secondary lower than 200 mOhm?
Capacitance (please note that this is tricky, as it is high in value, but limited to a small percetage of the winding, so being equivalent to a smaller capacitance)?
Primary inductance?
Other parameters?
Hi Roberto,
Sorry, but I did not understand your discussion of 3 kg and 5 kg cores above. I thought Toroidy was limited in their winding machines to the maximum core size as used on the KT88SE and PSE, which is really a 2 kg core.
It would be interesting to know what a Toroidy custom transformer for UNSET would cost as general info, but I suggest we build the prototype boards and see how our UNSET amps work with what we already have for output transformers. Then decide what is needed in custom design based on that experience. As for the design specs we may have different opinions about what is required.
For UNSET I am interested in sweep tubes that are characterized by high plate, low screen voltages, so no Ul tap is needed for me. I’m also fine with a single 8 Ohm secondary. I’m interested in CFB, but don’t know enough to know if it should be desirable in a custom transformer for UNSET.
Sorry, but I did not understand your discussion of 3 kg and 5 kg cores above. I thought Toroidy was limited in their winding machines to the maximum core size as used on the KT88SE and PSE, which is really a 2 kg core.
It would be interesting to know what a Toroidy custom transformer for UNSET would cost as general info, but I suggest we build the prototype boards and see how our UNSET amps work with what we already have for output transformers. Then decide what is needed in custom design based on that experience. As for the design specs we may have different opinions about what is required.
For UNSET I am interested in sweep tubes that are characterized by high plate, low screen voltages, so no Ul tap is needed for me. I’m also fine with a single 8 Ohm secondary. I’m interested in CFB, but don’t know enough to know if it should be desirable in a custom transformer for UNSET.
Last edited:
George, I have a pair of the Toroidy KT88SE 3k primary transformers, that I plan to use in my UNSET build. I’m willing to ship them to you for comparative testing (NOT destructive testing
). If you are interested and in a position to complete the testing in the next few weeks we could work that out in PM. The USPS fixed rate shipping should not be too expensive between us.
Progress with the Beta Board anyone?
I’m back from my vacation and reviewed my previous work. Noticed that I forgot to solder in the pico fuses in the cathodes of the power tubes. Will fix that before finishing my Beta boards.
@spiggs, I wondered how your listening went, and if the hum abatement efforts were satisfactory.
I’m back from my vacation and reviewed my previous work. Noticed that I forgot to solder in the pico fuses in the cathodes of the power tubes. Will fix that before finishing my Beta boards.
@spiggs, I wondered how your listening went, and if the hum abatement efforts were satisfactory.
I am happy with how quiet the amp is now. I cannot hear anything through the speakers with my ear right up next to them. I am putting together one of Pete Millett's filament supply boards to run the driver tube heaters to replace the
wallwart and ebay regulator I have hooked up now.
My overall listening impressions are good. I brought the UNSET in and hooked it up to my main speakers in place of my triode wired EL34 SSE. I liked what I heard and think the effort to get this in a nice case will be well worth it. More punch and dynamics with the same SE sound.
Having said that I would like to figure out why my build does not seem to meet spec. REW consistently shows about 1.2% THD at 1W and hits 5% THD at ~9W. Since this is my first use of REW I am not 100% confident in the readings but on my Modulus 86 amp I do measure much lower distortion so it seems the setup is capable of lower readings. I can get the THD slightly lower by playing with the bias but nothing dramatic. I will be very interested to see how other UNSETs measure.
@spiggs,
Thanks for the update. Glad the hum issue was solved by the DC heater supply for the driver tubes.
The distortion performance of your UNSET is concerning! One would expect that the SE UNSET would show much more distortion than a good Modulus 86, but 5% at 9 watts is disappointing. What could be the cause? Drivers?
Your REW measurement was with 6KT6 and 6DQ5 tubes, right? IIRC your 6KT6 plate current was 2.5 ma, plate voltage 125Vdc, but don’t recall the screen voltage. Do you have any 12GN7’s to try. IIRC George achieved his “good” distortion measurements with that driver.
I’m not familiar with the 6KT6 tube and looked up it’s datasheet. See the “characteristic performance” graph attached below, with plate and screen voltage both at 125V. It appears that at 2.5 ma plate current the 6KT6 is in an undefined (unlinear?) region of operation. But if this low plate current of the 6KT6 (with 3 watt max plate dissipation) present a problem, the problem could be worse using 12GN7 a much “bigger” tube with 11 watt dissipation. Of course we don’t know what George’s 12GN7 plate current was running.
Hope George’s workshop reorg goes well, and that he would be back in full force soon. We need him.
Thanks for the update. Glad the hum issue was solved by the DC heater supply for the driver tubes.
The distortion performance of your UNSET is concerning! One would expect that the SE UNSET would show much more distortion than a good Modulus 86, but 5% at 9 watts is disappointing. What could be the cause? Drivers?
Your REW measurement was with 6KT6 and 6DQ5 tubes, right? IIRC your 6KT6 plate current was 2.5 ma, plate voltage 125Vdc, but don’t recall the screen voltage. Do you have any 12GN7’s to try. IIRC George achieved his “good” distortion measurements with that driver.
I’m not familiar with the 6KT6 tube and looked up it’s datasheet. See the “characteristic performance” graph attached below, with plate and screen voltage both at 125V. It appears that at 2.5 ma plate current the 6KT6 is in an undefined (unlinear?) region of operation. But if this low plate current of the 6KT6 (with 3 watt max plate dissipation) present a problem, the problem could be worse using 12GN7 a much “bigger” tube with 11 watt dissipation. Of course we don’t know what George’s 12GN7 plate current was running.
Hope George’s workshop reorg goes well, and that he would be back in full force soon. We need him.
Attachments
Not much time for my addiction ehmm hobby lately.
Anyways, the OPTs that I have available are too high Z for a typical UnSET.
Also, I haven't settled on an output tube yet: I've collected too many types over the years.
Just now I came across some 8068's.
Weird thing with a tiny cathode, big 35W anode and a fragile screen grid.
To be continued...
Hey Parafeed813,
Good to hear from you again! Sorry about work overload slowing you down with your UNSET. Any thing wrong with your EL36 and E130L as outputs?
You achieved 14 Watts at 90 ma with a humble EL36 clipping symmetrically - I’d say that’s excellent. Have you done any distortion measurements yet? Good luck with your further explorations.
BTW, check the price on 8068 tubes, they go for around $60-100! Apparently they are highly sought-after for Kepco HV supplies. Sell them and get some lower cost tube on George’s list for UNSET.
Good to hear from you again! Sorry about work overload slowing you down with your UNSET. Any thing wrong with your EL36 and E130L as outputs?
You achieved 14 Watts at 90 ma with a humble EL36 clipping symmetrically - I’d say that’s excellent. Have you done any distortion measurements yet? Good luck with your further explorations.
BTW, check the price on 8068 tubes, they go for around $60-100! Apparently they are highly sought-after for Kepco HV supplies. Sell them and get some lower cost tube on George’s list for UNSET.
I have tried 3 different driver tubes so far, 6KT6, 6J51P, and 6EJ7. All using 6DQ5 output tubes. Results are all about the same in REW. I don't have any 12GN7 tubes but will dig and see if I have anything else.