Been spending some time (days, not hours) cataloguing the thermionic, sand and driver stash. Already found several goodies I forgot about, e.g. an externally beaten Quad 303 that fired-up after a brief ramp-up on the Variac.
Already received the output and p/s cap from Mouser...yeah, sand!
OK just kidding, although low voltages will work better with four grandchildren under four, so there's that.
Downside is that the Chief Cook & Bottle Washer has seized on this opportunity to force an eBay purge of about a dozen guitar speakers. Rats.
Also found P-P OPTs, including pairs of both Hammond 1608s and Dynaco Z565s (for the ST and SCA-35 EL84 uL amps) amongst others. These in particular are practically begging for a little solder on those old wires.
Questions for George:
(1) are the SPP boards available? Looks like I need a few. Lord knows I'm not doing point-to-point at my age if I can avoid it.
(2) I noticed that on the Dynaco EL-84 amps, the SPLITTER plate and cathode resistors are slightly different values (cathode is about 22% larger, 27K vs 22K), which seems odd.
On the SPP, they are the same (I would say symmetrical), which seems logical. What gives? D. Hafler up to some tricks I can't fathom?
(3) I would like to use one board to make a Flexi-SPP, with off-board Octals for 6V6 and the like. Maybe an octal driver, maybe even 6L6etc family power tubes if I can concoct an easy way to switch OPTs. Any issues with that?
On the SSE with off-board octals for 6AV5s, I neatly tie-wrapped the wires, but didn't twist any, not even the fils. No hum, but probably bad practice.
What SHOULD I do? Just twist the fils, say 3 turns/inch? Twist the whole bundle? Just curious.
(4) What, no CCS on the driver cathode? I found a sleeve of 10M45s in the stash...I can do it off-board like I did on the spit & duct tape Rat Lab 6AV7.
(5) Coupling caps on SE amps seem to cluster around 0.22uF, while P-P amps all seem to use 0.1uF. Why is that? Only dealing with half a wave at a time?
Thanks in advance...
Already received the output and p/s cap from Mouser...yeah, sand!
OK just kidding, although low voltages will work better with four grandchildren under four, so there's that.
Downside is that the Chief Cook & Bottle Washer has seized on this opportunity to force an eBay purge of about a dozen guitar speakers. Rats.
Also found P-P OPTs, including pairs of both Hammond 1608s and Dynaco Z565s (for the ST and SCA-35 EL84 uL amps) amongst others. These in particular are practically begging for a little solder on those old wires.
Questions for George:
(1) are the SPP boards available? Looks like I need a few. Lord knows I'm not doing point-to-point at my age if I can avoid it.
(2) I noticed that on the Dynaco EL-84 amps, the SPLITTER plate and cathode resistors are slightly different values (cathode is about 22% larger, 27K vs 22K), which seems odd.
On the SPP, they are the same (I would say symmetrical), which seems logical. What gives? D. Hafler up to some tricks I can't fathom?
(3) I would like to use one board to make a Flexi-SPP, with off-board Octals for 6V6 and the like. Maybe an octal driver, maybe even 6L6etc family power tubes if I can concoct an easy way to switch OPTs. Any issues with that?
On the SSE with off-board octals for 6AV5s, I neatly tie-wrapped the wires, but didn't twist any, not even the fils. No hum, but probably bad practice.
What SHOULD I do? Just twist the fils, say 3 turns/inch? Twist the whole bundle? Just curious.
(4) What, no CCS on the driver cathode? I found a sleeve of 10M45s in the stash...I can do it off-board like I did on the spit & duct tape Rat Lab 6AV7.
(5) Coupling caps on SE amps seem to cluster around 0.22uF, while P-P amps all seem to use 0.1uF. Why is that? Only dealing with half a wave at a time?
Thanks in advance...
4) I think George has written that the high gain of the 6BQ5 negates the need for the CCS on the input triode.
5) The C of the coupling cap and R of the following grid resistor make a one pole high pass filter. So you need to also look at the R to make any conclusion. I don't have an SPP schematic handy, so don't know the nominal specified value.
A 6AQ5 SPP is on my to do list, but my recent "projects" have been so lengthy, my to do list is getting to be more like a before I croak list.
5) The C of the coupling cap and R of the following grid resistor make a one pole high pass filter. So you need to also look at the R to make any conclusion. I don't have an SPP schematic handy, so don't know the nominal specified value.
A 6AQ5 SPP is on my to do list, but my recent "projects" have been so lengthy, my to do list is getting to be more like a before I croak list.
Last edited:
"before I croak list" : I hear that, which is why I'm asking these questions in the middle of the night. "If I die before I wake..."
That's one of the reasons I like the TubeLab boards; I can solder on up in no time relative to tedious point-to-point with tag boards. Plus I use those little single pin screw terminals for things like cathode resistors, so can change values with screwdriver.
I forgot about the cap-grid resistor "network" I guess you'd call it.
Among plate/cathode follower splitters I found (ie "split-load phase splitter")
LEAK EL84: 1M grid and 0.22 caps.
Heathkit UA-1 & AA-30: 1M grid and 0.1 caps
Eico HF-81: 330K grids and 0.25 caps
Dynaco/SPP: 470K grid and 0.1 caps
Heathkit SA2: 470K grid and 0.05 caps
Back to the books to understand the relationship...
That's one of the reasons I like the TubeLab boards; I can solder on up in no time relative to tedious point-to-point with tag boards. Plus I use those little single pin screw terminals for things like cathode resistors, so can change values with screwdriver.
I forgot about the cap-grid resistor "network" I guess you'd call it.
Among plate/cathode follower splitters I found (ie "split-load phase splitter")
LEAK EL84: 1M grid and 0.22 caps.
Heathkit UA-1 & AA-30: 1M grid and 0.1 caps
Eico HF-81: 330K grids and 0.25 caps
Dynaco/SPP: 470K grid and 0.1 caps
Heathkit SA2: 470K grid and 0.05 caps
Back to the books to understand the relationship...
I have exactly 4 SPP boards left in stock. Board sales have dwindled to 3 or 4 per month and the SPP is third in sales out of my 4 boards. I may order more after I total up 2020 and decide what to do in 2021.
In theory, all the current flowing through the cathode of an "ideal" triode must also flow through the plate, so if cathode current = plate current balance should be achieved with equal resistors IF an equal load is presented by the next stage. There are no "ideal" triodes, and a few other factors may influence balance, most notably mismatched output tubes. I found that equal resistor values worked best in the amps that I built, but you could put a lower value resistor in one leg with a small trimpot in series with it and tweak for lowest distortion.
The OPT will see voltages up to twice the B+ voltage in normal operation. Many small switches will NOT eat this much voltage without fireworks. I used "appliance" switches in my SSE builds with good success, but they were only rated for 240 VAC. Some larger plug and socket approaches may be OK. I used "Cinch Jones" connectors or "Euro Strip" screw terminals in my amps and test boards. You can also use ordinary octal tube sockets and bases from a dead tube like those 6AV5's, but the bases can also be ordered from many tube stores.
I have done some off board tube stuff with the SSE and TSE, but I haven't tried it with the SPP. I prefer to twist the heater / filament wires, and space the other's out flat and as close to the chassis as possible. Keep the grid wire as far away as possible from the heater and plate.
The driver is the splitter in the SSE, and a CCS anywhere in the splitter WILL mess up the balance. It could be possible to use a CCS in the plate circuit of the input stage to boost the gain and lower the THD as done in the SSE. I have not tried it though.
I designed the TSE in 2003, and committed it to a PC board in 2004. I started selling blank boards via the web site in 2005. I asked on the web site what people wanted next and the response was "something simple." This led to the creation of the SSE, which has been my biggest selling board ever since.
After the SSE, I asked again what people wanted and the reply was a simplified push pull amp like SY's Red Light District or Morgan Jones' Bevois Valley. The SPP is similar to both of those, only on a PC board. It never sold well at all. It seems that an audio club, a show, or a web site will feature one and I'll get a dozen or more orders from the same area in a couple months, then it's back to 6 or 8 a year, so I haven't experimented a lot with it as I have done with other stuff.
The coupling cap to the output tube grid is a big trade off between several things. Every output tube will have a minimum resistance in the grid circuit spec. Violate this by using a larger grid resistor and you risk output tube runaway and red plate death. Make the grid resistor too small, and you give up gain. In most cases the grid resistor is set to the maximum spec given in the tube manual. The capacitor value is then chosen for the desired bass rolloff frequency.
In an amplifier without global feedback the capacitor can be made pretty large, limited only by how much bass the OPT can pass. Most Single Ended triode amps do not use feedback, and can use a larger coupling cap.
The SPP is usually run in pentode mode, and global negative feedback is usually used to reduce distortion and lower the output impedance. This feedback comes from the secondary side of the OPT. Smaller sized OPT's will exhibit too much phase shift at low frequencies which can make the amp unstable if these frequencies are passed to the OPT. The low frequency limit must be increased to something the OPT can deal with by reducing the value of the coupling cap.
My "bucket list" is way bigger than my bucket. I have realized that many things will never be done, so it's time to do some more trimming.
I made a push pull Universal Driver board that may make things a bit easier than adding a bunch of off board sockets to an SPP board. It is a driver circuit capable of pushing some big transmitter tubes to 250 watts, or feeding little 50C5 radio tubes just enough to make 15 watts. It uses a pair of 6CG7's or other 9 pin tubes to generate enough drive to feed almost any tube. There are mosfet followers on the board to drive nearly any grid in class A, AB1 or AB2.
I have built a breadboard with a pair of driver boards, and I use turret boards for the output tubes. White "euro strip" screw terminals are used for connecting the OPT's. This makes swapping sets of OPT's or output boards a 20 minute job with no or minimal soldering. It is discussed here:
Tubelab Universal Driver Board, 2015 version
There are pictures of the breadboard scattered throughout the thread with different tubes and transformers. The white "euro strips" can be seen for the power supply connections on the left side in posts #36 and 41. I added a set for swapping OPT's later. I got them at Radio Shack, but they are gone for good here. Surviving Radio Shack stores seem to only sell phones now.
The breadboard in operation can be seen here. The video was made with a $200 cell phone, so the audio isn't exactly HiFi.
75 watt per channel triode mode tube amp prototype - YouTube
In theory, all the current flowing through the cathode of an "ideal" triode must also flow through the plate, so if cathode current = plate current balance should be achieved with equal resistors IF an equal load is presented by the next stage. There are no "ideal" triodes, and a few other factors may influence balance, most notably mismatched output tubes. I found that equal resistor values worked best in the amps that I built, but you could put a lower value resistor in one leg with a small trimpot in series with it and tweak for lowest distortion.
The OPT will see voltages up to twice the B+ voltage in normal operation. Many small switches will NOT eat this much voltage without fireworks. I used "appliance" switches in my SSE builds with good success, but they were only rated for 240 VAC. Some larger plug and socket approaches may be OK. I used "Cinch Jones" connectors or "Euro Strip" screw terminals in my amps and test boards. You can also use ordinary octal tube sockets and bases from a dead tube like those 6AV5's, but the bases can also be ordered from many tube stores.
I have done some off board tube stuff with the SSE and TSE, but I haven't tried it with the SPP. I prefer to twist the heater / filament wires, and space the other's out flat and as close to the chassis as possible. Keep the grid wire as far away as possible from the heater and plate.
The driver is the splitter in the SSE, and a CCS anywhere in the splitter WILL mess up the balance. It could be possible to use a CCS in the plate circuit of the input stage to boost the gain and lower the THD as done in the SSE. I have not tried it though.
I designed the TSE in 2003, and committed it to a PC board in 2004. I started selling blank boards via the web site in 2005. I asked on the web site what people wanted next and the response was "something simple." This led to the creation of the SSE, which has been my biggest selling board ever since.
After the SSE, I asked again what people wanted and the reply was a simplified push pull amp like SY's Red Light District or Morgan Jones' Bevois Valley. The SPP is similar to both of those, only on a PC board. It never sold well at all. It seems that an audio club, a show, or a web site will feature one and I'll get a dozen or more orders from the same area in a couple months, then it's back to 6 or 8 a year, so I haven't experimented a lot with it as I have done with other stuff.
The coupling cap to the output tube grid is a big trade off between several things. Every output tube will have a minimum resistance in the grid circuit spec. Violate this by using a larger grid resistor and you risk output tube runaway and red plate death. Make the grid resistor too small, and you give up gain. In most cases the grid resistor is set to the maximum spec given in the tube manual. The capacitor value is then chosen for the desired bass rolloff frequency.
In an amplifier without global feedback the capacitor can be made pretty large, limited only by how much bass the OPT can pass. Most Single Ended triode amps do not use feedback, and can use a larger coupling cap.
The SPP is usually run in pentode mode, and global negative feedback is usually used to reduce distortion and lower the output impedance. This feedback comes from the secondary side of the OPT. Smaller sized OPT's will exhibit too much phase shift at low frequencies which can make the amp unstable if these frequencies are passed to the OPT. The low frequency limit must be increased to something the OPT can deal with by reducing the value of the coupling cap.
My "bucket list" is way bigger than my bucket. I have realized that many things will never be done, so it's time to do some more trimming.
I made a push pull Universal Driver board that may make things a bit easier than adding a bunch of off board sockets to an SPP board. It is a driver circuit capable of pushing some big transmitter tubes to 250 watts, or feeding little 50C5 radio tubes just enough to make 15 watts. It uses a pair of 6CG7's or other 9 pin tubes to generate enough drive to feed almost any tube. There are mosfet followers on the board to drive nearly any grid in class A, AB1 or AB2.
I have built a breadboard with a pair of driver boards, and I use turret boards for the output tubes. White "euro strip" screw terminals are used for connecting the OPT's. This makes swapping sets of OPT's or output boards a 20 minute job with no or minimal soldering. It is discussed here:
Tubelab Universal Driver Board, 2015 version
There are pictures of the breadboard scattered throughout the thread with different tubes and transformers. The white "euro strips" can be seen for the power supply connections on the left side in posts #36 and 41. I added a set for swapping OPT's later. I got them at Radio Shack, but they are gone for good here. Surviving Radio Shack stores seem to only sell phones now.
The breadboard in operation can be seen here. The video was made with a $200 cell phone, so the audio isn't exactly HiFi.
75 watt per channel triode mode tube amp prototype - YouTube
