You are correct in the picture you show as the winding are "rotated" in the OPTs relative to the PT. Generally speaking, trannies for one manufacturer, whether they be PTs or OPTs, will have the winding in the same direction relative to the base where they attach to the chassis. The sure way to know is take the end bell off one side of a PT and an OPT.
S.
S.
Yes, I have carefully confirmed with sellers and manufacturers the winding orientation for the transformers as I considered them. The power transformer is being custom wound, and I did pay attention to the orientation so it will be correct. I had to specify the orientation for the custom power transformer. It is being wound for 120VAC instead of 110VAC like those from China.
Also for your consideration.
The Hammond 1650G
35W rated , no UL Taps, Raa is 6K6 rather than 8K but that will be not make much difference. Cheaper then the 1650F and likely to give better low bass due to its higher max rating.
Cheers Ian
The Hammond 1650G
35W rated , no UL Taps, Raa is 6K6 rather than 8K but that will be not make much difference. Cheaper then the 1650F and likely to give better low bass due to its higher max rating.
Cheers Ian
I'm sort of stuck on these offerings from Primary Windings for some reason ... they seem to fit the requirements and the price is not bad. I'm keeping a list of alternatives though. The Primary Windings 25 watters come in various impedences if 8k isn't ideal:
https://primarywindings.com/push-pull-output-transformers/?1091_product_cat[0]=54
$267.17 USD per pair delivered to the USA.
https://primarywindings.com/push-pull-output-transformers/?1091_product_cat[0]=54
$267.17 USD per pair delivered to the USA.
You may be kidding, but that's exactly what I did after carrying my amp several times around the house!
Attachments
Yes, that one, assuming 8k is the ideal value. One schematic says 8k and one says 10k. I have been under the assumption that 8k is the best choice. I have no way to test the stock Chinese transformers. They are labeled 10k on the side, not 8k.
I always have liked amps with rack mount handles on the front. Some people don't like them. When I was hauling around a Threshold pure class A amp that I could barely lift, they came in handy. The Dynaco ST-35 is hard to move around because the circuit boards are exposed and are right on the outer edges, and hands pinch them when moving it if not careful.
NGO,
10K is a little high for EL84s. Anything between about 6.5K and 8K should be fine.
Now that transformers have been decided, what are you going to use for coupling caps? If this might be your ultimate EL84 are you going to splurge on something exotic?
10K is a little high for EL84s. Anything between about 6.5K and 8K should be fine.
Now that transformers have been decided, what are you going to use for coupling caps? If this might be your ultimate EL84 are you going to splurge on something exotic?
Others have noted that 10k probably isn't ideal. With Chinese transformers we never know for sure without testing. I think XrayTonyB did test his transformers on YouTube and found them to be close to 8k.
As for the rest of the build, I have not gotten beyond the chassis and transformers. The power transformer will take 3-6 weeks. Then ouput transformers. Then I have to find time to build it, and I have a lot of satin black paint work to do first. When the time comes next year, I'll create a build thread. It's a PC board amp with a very small chassis, so massive coupling caps are not really an option. The kit comes with fake WIMA caps (I assume fake). I plan to upgrade or replace many of the components in the kit, although they do sound just fine. I'll put top quality filter caps in it for example for the sake of known reliability and peace of mind. I don't want to deal with Chinese filter caps erupting after a few years. I ordered a stepped attenuator made from SMD resistors that many people seem to like. One part at a time, I'll get it together.
As for the rest of the build, I have not gotten beyond the chassis and transformers. The power transformer will take 3-6 weeks. Then ouput transformers. Then I have to find time to build it, and I have a lot of satin black paint work to do first. When the time comes next year, I'll create a build thread. It's a PC board amp with a very small chassis, so massive coupling caps are not really an option. The kit comes with fake WIMA caps (I assume fake). I plan to upgrade or replace many of the components in the kit, although they do sound just fine. I'll put top quality filter caps in it for example for the sake of known reliability and peace of mind. I don't want to deal with Chinese filter caps erupting after a few years. I ordered a stepped attenuator made from SMD resistors that many people seem to like. One part at a time, I'll get it together.
A pentode [wired] EL84 has a plate impedance, rp, of about 40k Ohms.
An Ultra Linear [wired] EL84 has a plate impedance, rp, of approximately 3.5k or 5.6k based on the Ultra Linear tap percent, 40% or 26%, respectively.
A Triode [wired] EL84 has a plate impedance, rp, of about 1.7k Ohms.
In Class A push pull, both plate rp's drive the output transformer. A 10k center tapped transformer has 1/2 turns from a plate winding to the center tap.
1/2 turns = 1/4 impedance = 2.5k. But both tube's rp's are on, and driving [essentially in parallel], so the load impedance each plate sees is 5k.
In Class AB push pull, the tubes are both on, then the pull tube is off, then both tubes are on, then the push tube is off.
A 10k center tapped transformer has 1/2 turns from a plate winding to the center tap.
1/2 turns = 1/4 impedance = 2.5k.
As a result, whenever Class AB has one tube off, the tube that is on sees a load of only 2.5k.
Class A, tube plate rp's see 5k all the time. Constant load.
Class AB, tube plate rp's see 5k, then 2.5k, then 5k again. 2:1 load variation.
In Class AB, with minimum load of 2.5k, you can see that a 40k rp, or a 3.5 to 5.6k rp, might have a hard time driving such a low load impedance.
That means, for the 40k plate rp we MUST use negative feedback to reduce the effective rp;
and for Ultra Linear, if we do not use negative feedback to reduse the plate rp, we will have a damping factor of less than Unity (< 1).
Sometimes, Ultra Linear without negative feedback can work OK for some loudspeakers.
Note: Ultra Linear is a form of negative feedback, but is not enough to decrease the effective plate rp enough for optimal performance.
As you can see, there are lots of tradeoffs. All of those output tube circuits can work very well if the rest of the amplifier circuitry takes care of the output stage difficulties.
Just my opinions and observations.
An Ultra Linear [wired] EL84 has a plate impedance, rp, of approximately 3.5k or 5.6k based on the Ultra Linear tap percent, 40% or 26%, respectively.
A Triode [wired] EL84 has a plate impedance, rp, of about 1.7k Ohms.
In Class A push pull, both plate rp's drive the output transformer. A 10k center tapped transformer has 1/2 turns from a plate winding to the center tap.
1/2 turns = 1/4 impedance = 2.5k. But both tube's rp's are on, and driving [essentially in parallel], so the load impedance each plate sees is 5k.
In Class AB push pull, the tubes are both on, then the pull tube is off, then both tubes are on, then the push tube is off.
A 10k center tapped transformer has 1/2 turns from a plate winding to the center tap.
1/2 turns = 1/4 impedance = 2.5k.
As a result, whenever Class AB has one tube off, the tube that is on sees a load of only 2.5k.
Class A, tube plate rp's see 5k all the time. Constant load.
Class AB, tube plate rp's see 5k, then 2.5k, then 5k again. 2:1 load variation.
In Class AB, with minimum load of 2.5k, you can see that a 40k rp, or a 3.5 to 5.6k rp, might have a hard time driving such a low load impedance.
That means, for the 40k plate rp we MUST use negative feedback to reduce the effective rp;
and for Ultra Linear, if we do not use negative feedback to reduse the plate rp, we will have a damping factor of less than Unity (< 1).
Sometimes, Ultra Linear without negative feedback can work OK for some loudspeakers.
Note: Ultra Linear is a form of negative feedback, but is not enough to decrease the effective plate rp enough for optimal performance.
As you can see, there are lots of tradeoffs. All of those output tube circuits can work very well if the rest of the amplifier circuitry takes care of the output stage difficulties.
Just my opinions and observations.
It's class AB pentode. No ultra-linear. Global negative feedback. It will take me time to digest all of that info, but thanks! Maybe I can find a tutorial online.
Last edited:
Is there any legitimate benefit that might have caused the designer to have deliberately chosen the value of 10k instead of 8k? Voltage is 340. I really like the sound of the amplifier and I don't want to change the sound at all. If I didn't like it, a lot, I wouldn't be building a second one.
Note: the schematic in Post # 92 Does have global negative feedback.
That is a real good thing to do, because the EL84 is in pentode mode (40k plate impedance, rp).
1. The designer of that amplifier knew that loudspeaker manufacturers rated their speakers as "8 Ohm" But they actually have minimum impedances of 5 Ohms.
2. The designer knew that there was a glut of 10k output transformers, so they were less expensive than the otherwise similar 8k transformers.
3. The designer wanted a little more output power, so he abused the EL84 with more B+ voltage than the tube was rated for.
At that extra voltage, the 10k gave more output power than the 8k.
4. Fill in the blanks ______________________ ____________________ with your own stories.
That is a real good thing to do, because the EL84 is in pentode mode (40k plate impedance, rp).
1. The designer of that amplifier knew that loudspeaker manufacturers rated their speakers as "8 Ohm" But they actually have minimum impedances of 5 Ohms.
2. The designer knew that there was a glut of 10k output transformers, so they were less expensive than the otherwise similar 8k transformers.
3. The designer wanted a little more output power, so he abused the EL84 with more B+ voltage than the tube was rated for.
At that extra voltage, the 10k gave more output power than the 8k.
4. Fill in the blanks ______________________ ____________________ with your own stories.
I'll select #1 and #3. 😀
Whatever they did, it sounds great.
#3 had escaped my attention. I never looked at it. B+ is 340V. About 330V at the tube I think. I have it written down somewhere. Tube max rating is 400V. Does the voltage swing put it over 400V? I'm not going to change it. I have a few sets of 6P14P-ER (Russian 6N14N-EP) tubes that sound great, and they are rugged with 14W dissipation and 10,000 hr life ratings. Old Soviet-era military tubes.
Last edited:
Some EL84 variants are rated as high as 400 volts on the pate. But, most EL84, whether plain vanilla or fancy flavours, won't tolerate much more than 300 volts on their screen grids.
I've had good experiences buying from this fellow in Ukraine: https://nextfive.pro/
I believe he sells on eBay but also has an online store. The stock at the online store isn't fully up to date and shows zero tubes.
However, he has some decent stock and will send you a list of these and other parts if you drop him an email at: info@nextfive.pro
There are NOS EL84 and other desirable audio tubes at very fair prices. Just ask.
No association etc.
Steve
I believe he sells on eBay but also has an online store. The stock at the online store isn't fully up to date and shows zero tubes.
However, he has some decent stock and will send you a list of these and other parts if you drop him an email at: info@nextfive.pro
There are NOS EL84 and other desirable audio tubes at very fair prices. Just ask.
No association etc.
Steve
The amp has a regulated screen supply. I attributed this to the Chinese 6P14 that is reported to almost instantly self destruct if it goes over 300V. Mine is at 297V and I am not using the Chinese 6P14. I had to change one of the Zener diodes to get it below 300. One comes in the kit. I don't what makes the amp sound good, but it does. It gets the "timing" of the music right. Not sure how else to put it. I want to see what bigger better OPTs bring to the table. Thus, this thread looking for OPTs.
Soviet 6P14P-ER specs (service life >10,000 hours)
6P14P same but service life 3,000 hours
6P14P-EV same but service life 5,000 hours
Soviet 6P14P-ER specs (service life >10,000 hours)
6P14P same but service life 3,000 hours
6P14P-EV same but service life 5,000 hours
Last edited:
Thanks for the link and email address. I bought a quad of 6P14P-ER from Ukraine a while back. Very good. I bought another quad from Khazakstan for the ST-35, and they were good, so I ordered several more quads from the same person. I also have a quad from a US audiophile who bought a bulk lot of 500 6P14P-ER tubes and matched them. He kept a lifetime supply for himself and sold the rest, likely at a decent profit.
I looked at some of the EL84 data sheets at Franks tube data site.
EL84 Maximum Plate Voltage:
Quiescent 300V
Cut off (no current) 550V
The screen voltages are the same as the plate:
Quiescent 300V
Cut off (no current) 550V
That means you can triode wire an EL84 and use 300V Quiescent.
Using an EL84 beyond its ratings is possible, but . . .
Your Mileage May Vary.
There is a tube that is similar that is rated for higher quiescent plate voltage.
I am sure someone will remember the tube type number (4 numbers, in the thousands).
Sometimes I think I would like to try a pair of SV83 tubes. The maximum screen voltage is only200V.
Might be nice in Pentode mode, with an OD3 150V gas regulator tube for the screen.
EL84 Maximum Plate Voltage:
Quiescent 300V
Cut off (no current) 550V
The screen voltages are the same as the plate:
Quiescent 300V
Cut off (no current) 550V
That means you can triode wire an EL84 and use 300V Quiescent.
Using an EL84 beyond its ratings is possible, but . . .
Your Mileage May Vary.
There is a tube that is similar that is rated for higher quiescent plate voltage.
I am sure someone will remember the tube type number (4 numbers, in the thousands).
Sometimes I think I would like to try a pair of SV83 tubes. The maximum screen voltage is only200V.
Might be nice in Pentode mode, with an OD3 150V gas regulator tube for the screen.
Last edited:
Contact Ken Loo at J&K Audio. Ken does terrific work, will custom wind to your spec, probably also has an existing design that meets your requirements . Very high quality product, fair pricing.
Ken wound me a pair of 8k anode to anode with four individual secondary windings to configure how one wishes, dual 10% cathode feedback windings and UL taps with M3 lams. USD350 ex works Penang in January 2022. 20W at 20Hz to 20kHz.
A small air gap to accommodate some primary imbalance. I spec’d 80mA per primary winding to have some versatility.
I’ve used these with EL84s, they sound great.
None of this namby pamby 30Hz max power rating.
He also does M6 lams (lower price) C-core in GOSS, amorphous and finemet cores as well.
If you prefer OCC magnet wire, that can be accommodated. Silver magnet wire as well if you are so inclined.
Google jandkaudio or FB.
I am merely a satisfied customer, nothing more, nothing less.
Last edited: