Is this in AB, or strict class A?
The speakers I currently use (karlsonator with dual 16 ohm faital pro 3FE25 wired in parallel for 8 ohm) can also use another set of drivers I have ( peerless TC9FD1808, 8 ohm) wired in parallel for 4 ohm. Would this be a bit better match since its a 3k load at that point? I've been wanting to pull a pair of the 3FE25 out to try in another project anyway, so it wouldn't be a big deal at all to swap over.
With 250V, 60mA, 6000 Ra-a, it's all Class A. With 300 Ra-a, it's Class AB with a limit of 2.2W class A. The total power barely goes up.
With 320V, 60mA, 6000 Ra-a, you get 4.5W watts, still all class A, and 3000 Ra-a will give you 5.25W, with a 2W class A limit.
Interestingly, if you use a pair of VPT12-4170 in series interleave, you'd get almost 15W with a 2.5W class A limit. Higher efficiency and lower DCR...
Did you say you have 6P43P? You'll get about 20% more power (with 6k load) from them than the 12AV5GA, and lower THD, too. If you go 3k load, they will make 8W, but 22W with the VPT12s (2k2 load)... 320V B+, 35mA OP.
Keep in mind these are only calculations, and I find in reality there is usually a mismatch.
With 320V, 60mA, 6000 Ra-a, you get 4.5W watts, still all class A, and 3000 Ra-a will give you 5.25W, with a 2W class A limit.
Interestingly, if you use a pair of VPT12-4170 in series interleave, you'd get almost 15W with a 2.5W class A limit. Higher efficiency and lower DCR...
Did you say you have 6P43P? You'll get about 20% more power (with 6k load) from them than the 12AV5GA, and lower THD, too. If you go 3k load, they will make 8W, but 22W with the VPT12s (2k2 load)... 320V B+, 35mA OP.
Keep in mind these are only calculations, and I find in reality there is usually a mismatch.
Last edited:
Hmmmm 
I need to play with the power transformers I have on hand and see if I can't get more voltage out of them. They are industrial control types with a single multiple tap primary (480, 277, 240, 208, 0) and a 120 volt secondary. I'm running them backwards off of the the 277 volt secondary but with winding losses from the configuration I'm only getting ~270 volts after the rectifier. I'll try to maybe run from the 480-208 taps instead and see if the loss is any better. Maybe I can try a doubler off of the 0/208 windings even
I already have an EL86 pentode connected (my daily use one, even) amplifier that I plan to use with 6P43P triode connected once this project is finished, so I'll try to get something together that's reasonable here to use the 12AV5GA if at all possible.
If I had a hundred bucks to spend on iron I would be making an order from Antek right now
I need to play with the power transformers I have on hand and see if I can't get more voltage out of them. They are industrial control types with a single multiple tap primary (480, 277, 240, 208, 0) and a 120 volt secondary. I'm running them backwards off of the the 277 volt secondary but with winding losses from the configuration I'm only getting ~270 volts after the rectifier. I'll try to maybe run from the 480-208 taps instead and see if the loss is any better. Maybe I can try a doubler off of the 0/208 windings even
I already have an EL86 pentode connected (my daily use one, even) amplifier that I plan to use with 6P43P triode connected once this project is finished, so I'll try to get something together that's reasonable here to use the 12AV5GA if at all possible.
If I had a hundred bucks to spend on iron I would be making an order from Antek right now
Last edited:
I may have to try a pair of those at some point. Look like a quad wouldn't be too pricy from Arrow, free shipping over $50USD, and $17.33USD each-
https://www.arrow.com/en/products/vpt12-2080/triad-magnetics?gclid=EAIaIQobChMIk4yR753-5QIV4BitBh3yaQANEAYYASABEgL8Z_D_BwE
I was considering these originally before money got tight again, wonder how they compare? 1.1 pounds, versus 14 ounces, otherwise they look pretty similar.
AN-0206 - 25VA 6V Transformer - AnTek Products Corp
Would be a good match with one of these power transformers per channel (secondaries in parallel) which is a good value at $28USD. All the iron for a pair of monoblocks for $91USD shipped isn't too bad at all. If building all on a single chassis you could go with a single 100VA like this one and you would even save ~$22. Shipping is pretty reasonable from Antek within the US, even for the four together it's only $13.50 priority mail.
I use a set of these on the EL86 amplifier with series connected secondaries, for roughly 4200k into 8 ohms, and have been very happy with them. They would probably have been a good choice too, for a pentode connected build at least. Maybe overkill on triode mode?
As it was I was only able to afford the pair of enclosures for this build with the birthday money I got in a card this year
https://www.arrow.com/en/products/vpt12-2080/triad-magnetics?gclid=EAIaIQobChMIk4yR753-5QIV4BitBh3yaQANEAYYASABEgL8Z_D_BwE
I was considering these originally before money got tight again, wonder how they compare? 1.1 pounds, versus 14 ounces, otherwise they look pretty similar.
AN-0206 - 25VA 6V Transformer - AnTek Products Corp
Would be a good match with one of these power transformers per channel (secondaries in parallel) which is a good value at $28USD. All the iron for a pair of monoblocks for $91USD shipped isn't too bad at all. If building all on a single chassis you could go with a single 100VA like this one and you would even save ~$22. Shipping is pretty reasonable from Antek within the US, even for the four together it's only $13.50 priority mail.
I use a set of these on the EL86 amplifier with series connected secondaries, for roughly 4200k into 8 ohms, and have been very happy with them. They would probably have been a good choice too, for a pentode connected build at least. Maybe overkill on triode mode?
As it was I was only able to afford the pair of enclosures for this build with the birthday money I got in a card this year
I run the B+ for the just the 4 6P43Ps from a 100VA transformer, and it's at it's limit.
I've never tried the Antek because of the costs to Canada being crazy compared to Triad from Mouser of Digi-key but I'll bet they will be fine.
The higher the VA rating, the lower the DCR, and higher the efficiency. On low current it doesn't matter much, but when you try and push 1A through a 100R winding, if doesn't work very well (you lose power).
I use a pair of 25VA coils for 6P43P, a pair of 50VA coils will be used to build something with 12AV5GA I think, That, 6BQ6, or 6P36S, a pair or 100VA coils for the 36LW6 amp, and a pair of 250VA for the PPP 6P45S monoblocs. By the time you get to the 250VA coils, your paying almost as much as for Hammond, but the DCR is lower, and there are more choices for a low Z primary. If I build the monoblocs again, I'd use the 160VA version I think.
I've never tried the Antek because of the costs to Canada being crazy compared to Triad from Mouser of Digi-key but I'll bet they will be fine.
The higher the VA rating, the lower the DCR, and higher the efficiency. On low current it doesn't matter much, but when you try and push 1A through a 100R winding, if doesn't work very well (you lose power).
I use a pair of 25VA coils for 6P43P, a pair of 50VA coils will be used to build something with 12AV5GA I think, That, 6BQ6, or 6P36S, a pair or 100VA coils for the 36LW6 amp, and a pair of 250VA for the PPP 6P45S monoblocs. By the time you get to the 250VA coils, your paying almost as much as for Hammond, but the DCR is lower, and there are more choices for a low Z primary. If I build the monoblocs again, I'd use the 160VA version I think.
The Antek AS-1T275 that Lingwendil referenced is rated at 180 ma for each of its two HV secondaries. What are you averaging in your 6P43Ps?
Thought you might find this interesting.
Ran the sims for 36LW6 trioded, B+ 300V, 80mA OP.
With VPT18-5560 in series (100VA) Pout=50W
With VPT18-13800 in series (250VA) Pout=75W.
The only difference is the DCR and regulation spec. VPT18-5560 primaries are 45R, VPT18-13800 is only 6R... That makes a 25W difference apparently.
Ran the sims for 36LW6 trioded, B+ 300V, 80mA OP.
With VPT18-5560 in series (100VA) Pout=50W
With VPT18-13800 in series (250VA) Pout=75W.
The only difference is the DCR and regulation spec. VPT18-5560 primaries are 45R, VPT18-13800 is only 6R... That makes a 25W difference apparently.
I wish I had the 160VA version to measure the DCR... I'll send a message to Triad and ask them for the information. The 160VA is the "price point", it's 10$ more than the 100VA, but 40$ less than the 250VA Still, you build them once. What's another 160$ for another 20-30W of power and lower THD?
Still, you build them once. What's another 160$ for another 20-30W of power and lower THD?
Very cool. I've had my eye out for *LW6 tubes for a while, but never find them in cheap lots. I've got a 150VA dual primary, dual secondary (120:120 both sides) toroid that I've been holding onto for a big project that I'd love to use for a build with something like them. That's some serious power in triode for just two tubes!
So I did more testing last night, and the 208 winding gives 175VAC under load (well, with the 0,277 volt taps rectified and loaded) and curiously, I'm getting only 280VDC rectified with a 47uF cap, and a 2k resistor array load (140mA draw) from the AC voltage of 240. This sounds way too low!? The transformer runs barely even warm, in fact it doesn't feel more than maybe a couple degrees above ambient at this draw. I tried both with a diode bridge made from 1N4007 and a bridge rectifier (RS405L) and get the same voltages. This seems a little weird to me, as I would expect a bit closer to 320~330VDC starting from 240VAC. Any ideas why it's so low? This is a little weird in my opinion. I could just use a voltage doubler or something to get higher voltage.
Alternative idea- I could do a voltage doubler arrangement- I would ground the 240V tap, put a bridge across the 0,480 taps, and then capacitor-couple a bridge to this, stacking that bridges' voltage on top of the positive output. This would give me a negative rail for bias, and two positive rails (I would only use the higher one, at least for a triode arrangement, but it gives future options for pentode screen supplies, source followers, etc) so should work nicely. Like this-
I have plenty of suitable capacitors for this, and I could easily account for sag with the mosfet filters I'll be using since they work best with a bit of extra headroom to shave off crud anyway. I just want to at least understand the loss of AC during rectification before I get into any exotic/less usual rectification schemes.
So I did more testing last night, and the 208 winding gives 175VAC under load (well, with the 0,277 volt taps rectified and loaded) and curiously, I'm getting only 280VDC rectified with a 47uF cap, and a 2k resistor array load (140mA draw) from the AC voltage of 240. This sounds way too low!? The transformer runs barely even warm, in fact it doesn't feel more than maybe a couple degrees above ambient at this draw. I tried both with a diode bridge made from 1N4007 and a bridge rectifier (RS405L) and get the same voltages. This seems a little weird to me, as I would expect a bit closer to 320~330VDC starting from 240VAC. Any ideas why it's so low? This is a little weird in my opinion. I could just use a voltage doubler or something to get higher voltage.
Alternative idea- I could do a voltage doubler arrangement- I would ground the 240V tap, put a bridge across the 0,480 taps, and then capacitor-couple a bridge to this, stacking that bridges' voltage on top of the positive output. This would give me a negative rail for bias, and two positive rails (I would only use the higher one, at least for a triode arrangement, but it gives future options for pentode screen supplies, source followers, etc) so should work nicely. Like this-
I have plenty of suitable capacitors for this, and I could easily account for sag with the mosfet filters I'll be using since they work best with a bit of extra headroom to shave off crud anyway. I just want to at least understand the loss of AC during rectification before I get into any exotic/less usual rectification schemes.
Attachments
Ya that IS weird. Maybe the DCR is high or something...
On another note, one of the 36LW6 tubes seems to have a bad screen now... No current flow anymore, but the 1/4W 220R screen resistor didn't blow. Now I'll convert it to 6P36S. since a quad of them is the same as the cost of 1 36LW6. I also found out that my PP design will work SE if you remove one of the tubes. Apparently 100mA isn't enough for core saturation LOL
On another note, one of the 36LW6 tubes seems to have a bad screen now... No current flow anymore, but the 1/4W 220R screen resistor didn't blow. Now I'll convert it to 6P36S. since a quad of them is the same as the cost of 1 36LW6. I also found out that my PP design will work SE if you remove one of the tubes. Apparently 100mA isn't enough for core saturation LOL
Well, trimmed the excess metal from the power transformers, and started drilling the enclosures, as that's the number one thing that keeps me from finishing projects. Having these ready will make less work for me anyways. Got all the holes drilled except for the power transformers and the IEC jacks. I'll need to grab some grommets for where the wires pass through the top, and I'm thinking stainless screws all around, maybe brass
Attachments
I'm coming in on this one a bit late. The Williamson, unless built similar to the original is known to be unstable under NFB conditions. Keep in mind the original was built with triode output tubes, their gain varies very little driving into a complex load such as a loudspeaker. Not so the pentode output stage. So what might be 20 db using a test resister as a load could become 35-40 db at some frequencies with a loudspeaker load.
Williamson used a very good OPT, reasonably free of leakage inductance & stray wdg capacity. It was quite expensive. That kind of OPT doesn't come cheap even today.
Still possible to build a la Williamson but great care & planning are needed with the NFB. Careful BW limiting at both ends of the spectrum is required. Don't try for DC to Green Light!!
PS- I never built a Williamson, everyone else did that. I favored exploring other ways. But some of the guys in the lab I worked in circa 1956 did. It sounded OK on an electrostatic speaker. Also built in the lab using Saran Wrap & 4x8 ft sheets of prepunched masonite. A major project in itself. Even with that strange load their Williamson was stable.
Williamson used a very good OPT, reasonably free of leakage inductance & stray wdg capacity. It was quite expensive. That kind of OPT doesn't come cheap even today.
Still possible to build a la Williamson but great care & planning are needed with the NFB. Careful BW limiting at both ends of the spectrum is required. Don't try for DC to Green Light!!
PS- I never built a Williamson, everyone else did that. I favored exploring other ways. But some of the guys in the lab I worked in circa 1956 did. It sounded OK on an electrostatic speaker. Also built in the lab using Saran Wrap & 4x8 ft sheets of prepunched masonite. A major project in itself. Even with that strange load their Williamson was stable.
I'm going triode mode here (for once!) so worst case I may forgo NFB to the first stage if I can't get a stable output. 6K plate to plate should still get me a reasonable enough damping factor, but the DCR of my output transformer may not help things. I've left room to swap out for better output iron later on as budget allows, and I'm considering balanced feedback from a center tapped secondary to the cathodes of the drivers if I go that way eventually. I need to experiment with that a bit soon.
Maybe it would be worth looking into plate-to-grid feedback in the output stage? The 6N1P should be able to swing into it.
Still possible to build a la Williamson but great care & planning are needed with the NFB. Careful BW limiting at both ends of the spectrum is required. Don't try for DC to Green Light!!
I'll toss in a bit of compensation at the plate of the first triode, and across the NFB resistor for sure.
That sounds like a neat speaker, I've not looked into ESLs much but they sound very interesting.
My last williamson used this iron, 6SL7 input/concertina, 6SN7 differential, and 6AS7G output tubes into 4 ohm speakers. It sounded great and needed just a bit of capacitance across the feedback resistor to pretty up the square wave response. Hopefully this will behave as well.
Last edited:
As I said, REALLY interesting. I have yet to learn Spice, but I wonder how reliable the tube models for big sweep tube and sims are. I got so say, Koda, I really enjoy you tube adventures recorded here. Many thanks.
Little off topic, but I got a clutch of NOS 26LW6’s I bought years ago, and have since wondered what I could do with them. Hmmm! Evil thoughts are popping. Evil because sweet wifey wants me to complete several projects before starting another one.
I couldn't really get the hang of spice for whatever reason
Honestly much of why I still build is because of koda's posts. He got me back into tubes and is a great guy to bounce ideas off of and has some very cool builds.
BUILD THEM INTO A MONSTER
At the very least, start a shoebox full of parts to go with them
Before you know it you'll just need to buy a chassis and some iron
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.