since arnold is inclined to use IT's, then this is a good read: http://www.izzy-wizzy.com/audio/powamp813.html
Hi Tony, not IT all the way 😀 because I only have one pair of 1660S and I'm just trying to 'salvage' it. That's a nice link by the way, and I have to read it. 🙂
Re: What's cookin'
Since the LTP is capacitor coupled, you don't need that resistor with active tail loading. I've done two different LTPs (6SL7s and cascoded 6BQ7As) and have gotten excellent AC balance.
If you're doing direct coupling, the DC offset adjust resistor is usually in the cathode circuit, not the plate since the voltage at the cathode is much lower, and it doesn't affect AC balance nearly as much.
arnoldc said:
Since the LTP is capacitor coupled, you don't need that resistor with active tail loading. I've done two different LTPs (6SL7s and cascoded 6BQ7As) and have gotten excellent AC balance.
If you're doing direct coupling, the DC offset adjust resistor is usually in the cathode circuit, not the plate since the voltage at the cathode is much lower, and it doesn't affect AC balance nearly as much.
Corrrect. See Pete Millett's site for a similar 813 SET design. Also download his .pdf of pentode/triode tests.Miles Prower said:
Doesn't need the second stage. I ran with pentode 12HL7/12HG7 as a single stage drive to the 813's. I had enough gain on HE speakers.resident said:
that's cool Brett! I checked Pete Millet's website and it's a nice pentode! I'll try it in my 2A3SET. Now, I use SV83 with great success...
Seems like a good idea, an LTP with pentodes to drive 300Bs or 2A3s in P-P ... it can be used as a universal input stage for a lot of P-P experiments.
Seems like a good idea, an LTP with pentodes to drive 300Bs or 2A3s in P-P ... it can be used as a universal input stage for a lot of P-P experiments.
hey-hey!!!,
I have a 2A3 PP amp drawn. It uses a 6BL7GTA is a LTP like the previous 12AT7 schematic. It runs a high voltage B+ with a Blumlein Garter balancing/bias. This burns twice the usual cathode bias, but delivers very nice idle current match. If the pdf isnt too big, it will appear below. If it doesn't send me an email and perhaps somebody can modify it...🙂 ( it is too big)
The high B+ offers a substantial headroom to the 6BL7 stage, but no positive grid drive is available 'cause it is cap coupled.
cheers,
Douglas
I have a 2A3 PP amp drawn. It uses a 6BL7GTA is a LTP like the previous 12AT7 schematic. It runs a high voltage B+ with a Blumlein Garter balancing/bias. This burns twice the usual cathode bias, but delivers very nice idle current match. If the pdf isnt too big, it will appear below. If it doesn't send me an email and perhaps somebody can modify it...🙂 ( it is too big)
The high B+ offers a substantial headroom to the 6BL7 stage, but no positive grid drive is available 'cause it is cap coupled.
cheers,
Douglas
Re: What's cookin'
The driver tubes are going to be an issue here, the Rp is too high and that's going to do bad things to the bandwidth. Use something like a 6BL7, lower Rp and can handle a fair amount of current, you'll have much better bandwidth and a cleaner sound as well.
The 56/76 are great tubes, I'm using them as the input tubes in my headphone amp, but I think that having them drive a 300B through an interstage transformer is a bit too much to ask.
arnoldc said:
The driver tubes are going to be an issue here, the Rp is too high and that's going to do bad things to the bandwidth. Use something like a 6BL7, lower Rp and can handle a fair amount of current, you'll have much better bandwidth and a cleaner sound as well.
The 56/76 are great tubes, I'm using them as the input tubes in my headphone amp, but I think that having them drive a 300B through an interstage transformer is a bit too much to ask.
Re: Re: What's cookin'
Yeah, I would say so too, and FWIW I think my old design compromises bandwidth due to miller capacitance in the output tubes. If you use George's power drive with this circuit or a good CF you can get at least 80kHz driver stage bandwidth with my design and probably comparable with 56/76. If you use an interstage consider some of the higher transconductance triode or triode connected pentodes out there.
Kevin K
aerius said:
Yeah, I would say so too, and FWIW I think my old design compromises bandwidth due to miller capacitance in the output tubes. If you use George's power drive with this circuit or a good CF you can get at least 80kHz driver stage bandwidth with my design and probably comparable with 56/76. If you use an interstage consider some of the higher transconductance triode or triode connected pentodes out there.
Kevin K
Hi Douglas, I'd been looking at the Blumlein Garter cct for some old guitar amps I have here on the bench. Good to hear it works as well as JB predicted. Nice simple solution for cathode biassed stages, and much easier/cheaper than the fixed bias regulator he published around the same time.Bandersnatch said:
It works very well and has lowish output Z too. These eventually replaced all sorts of previous driver ccts including some very nice two stage tride designs with and without IT's (on the 813PP).resident said:
Funny you shoutls mention that. I'm building a new workshop setup and want to construct some 'modules' to make experimentation easy similar to Tubelab's ones. The 12HL7's in both pentode and triode will be part of that. Add a few dozen clipleads and we have amps galore.resident said:Seems like a good idea, an LTP with pentodes to drive 300Bs or 2A3s in P-P ... it can be used as a universal input stage for a lot of P-P experiments.
The use of individual bias adjustments for each 300B, which I think is very important
Could this be done better?
Assume the current though the 330R (Rk of 300B) is 65mA. Vk=21.45V. I would like to use LM317 CCS to replace the 330R and set to 65mA. The -V supply should connect to the centre tap or IT's secondary and adjust the -V, so that the V drop accross LM317 is about 22V. The DC blocking Caps. and grid leak resistors are no needed. Not sure the bypass cap is needed on the LM317?
Could this be done better?
Assume the current though the 330R (Rk of 300B) is 65mA. Vk=21.45V. I would like to use LM317 CCS to replace the 330R and set to 65mA. The -V supply should connect to the centre tap or IT's secondary and adjust the -V, so that the V drop accross LM317 is about 22V. The DC blocking Caps. and grid leak resistors are no needed. Not sure the bypass cap is needed on the LM317?
Im trying to build a workshop setup too. But not too much time to do everything... Now I m building a universal SE set-up. The regulated PSU is ready and now need to build the amp.Brett said:
The idea of using 12HL7 or 12HG7 was in the right time. 🙂 thanks!
By the way, have you heard about the tetrode 6E5P? A fellow in an other forum use it as a driver.... I didn't have time to read all the articles...
Mount a stopper resistor right at the grid pin on the tube socket, otherwise you will have oscillators, transmitters and TV jammers galore. 12HL7, 12HG7, 12BY7, 12GN7, 6AG7, 6AC7 and 6CL6 are all video output tubes, they don't need much provocation to oscillate. I learned that one the hard way.
hey-Hey!!!,
For pentode LTP, there's a few considerations. First is the screen grids need bypassed to the cathode node. Dropping R from B+ to set proper votlage and a cap to cathodes.
Of particular importance is the grounded grid be rigged with a stopper. In the LTP, that phase is running as a cathode driven stage and those are quite oscillation prone. Make it a 12GN7 and you're loaded for Cape Buffalo.
The pentode delivers a fairly steady output Z, based on its plate load, not varying like a triode which is dominated by a varying plate Z.
Making a LTP out of a cascode is even better than pentodes. The current path is a series one. What leaves the common cathode node will leave by one of the plate loads. Pentodes are plagued by the g2 path. One can get more headroom, and make a better cascode( higher output Z ), by stacking a MOSFET on top instead of another triode. The part of the plate curves to the left of 2x the upper grid voltage I refer to as, 'here be Monsters' only extends a little past the gate voltage.
cheers,
Douglas
For pentode LTP, there's a few considerations. First is the screen grids need bypassed to the cathode node. Dropping R from B+ to set proper votlage and a cap to cathodes.
Of particular importance is the grounded grid be rigged with a stopper. In the LTP, that phase is running as a cathode driven stage and those are quite oscillation prone. Make it a 12GN7 and you're loaded for Cape Buffalo.
The pentode delivers a fairly steady output Z, based on its plate load, not varying like a triode which is dominated by a varying plate Z.
Making a LTP out of a cascode is even better than pentodes. The current path is a series one. What leaves the common cathode node will leave by one of the plate loads. Pentodes are plagued by the g2 path. One can get more headroom, and make a better cascode( higher output Z ), by stacking a MOSFET on top instead of another triode. The part of the plate curves to the left of 2x the upper grid voltage I refer to as, 'here be Monsters' only extends a little past the gate voltage.
cheers,
Douglas
Last night I calculated what I needed for the input differential stage using the WE396A-
It will have -12V on the tail, 7K5 plate resistors, off a 160V B+ which would give me Av=18 and Vout of +-60V
It will be direct coupled to the next stage-
The second differential stage using 76 will be biased to some -15V and with 127 into its grid needs 142V on the cathode, so my negative supply needs to be that high. It will have 18K on the plate fed from 300V B+ which would give me Av=8 and Vout of +-130V
After reading the replies, I think I'm back to the drawing board. If I ditch the LL1660S I think my life will be much simpler or use the WE396A to drive it and use 5687 as the input differential.
😕 😕 😕
It will have -12V on the tail, 7K5 plate resistors, off a 160V B+ which would give me Av=18 and Vout of +-60V
It will be direct coupled to the next stage-
The second differential stage using 76 will be biased to some -15V and with 127 into its grid needs 142V on the cathode, so my negative supply needs to be that high. It will have 18K on the plate fed from 300V B+ which would give me Av=8 and Vout of +-130V
After reading the replies, I think I'm back to the drawing board. If I ditch the LL1660S I think my life will be much simpler or use the WE396A to drive it and use 5687 as the input differential.
😕 😕 😕
hi arnold,
how about your output stage B+ and output traffo specs? knowing those can help determine grid bias and drive requirements.
you do not have to ditch your lundhall if you want it...
how about your output stage B+ and output traffo specs? knowing those can help determine grid bias and drive requirements.
you do not have to ditch your lundhall if you want it...
Hi Douglas
I remember you wrote about the triode/mosfet cascode before, run from the UL taps of the output transformer. I think it was in gingertube's baby Huey thread.
Do you have some schematic available on the proposed triode/mosfet cascode? And measurements (distortion, gain, etc)?
Many thanks, Erik
I remember you wrote about the triode/mosfet cascode before, run from the UL taps of the output transformer. I think it was in gingertube's baby Huey thread.
Do you have some schematic available on the proposed triode/mosfet cascode? And measurements (distortion, gain, etc)?
Many thanks, Erik
hey-Hey!!!,
I have not made those measurements. Gain is gm-dependant on the plate load. Gain for me is not always gain for you if the OP is different. I don't yet have distortion measurement devices, so that happens only seldom when I can arrange a loan.
cheers,
Douglas
I have not made those measurements. Gain is gm-dependant on the plate load. Gain for me is not always gain for you if the OP is different. I don't yet have distortion measurement devices, so that happens only seldom when I can arrange a loan.
cheers,
Douglas
- Status
- Not open for further replies.