I'm finally (FINALLY!) getting started on a minimalist Baby Huey EL84 build. I was going to try EL86s, but since I have the iron from a Dynaco ST35, I figured I should use it. The problem is, its power transformer (PA-774) has a secondary rated at 330V-0-330V at 180mA. Full-wave rectification using a couple of UF4007 silicon diodes will probably yield 420V DC. That's an awfully high voltage for EL84s, even with cathode bias.
My question is, has anyone used this PA-774 transformer in a Baby Huey (or similar) build, using today's EL84 tubes? Did you get it to work without burning up your EL84s? If so, how did you do it?
I'm thinking of fitting a CLC after the UF4007s, then a second RC stage consisting of a 270R 15W resistor and 470uF 450V capacitor. But it sure is wasteful. Seriously wasteful.
My question is, has anyone used this PA-774 transformer in a Baby Huey (or similar) build, using today's EL84 tubes? Did you get it to work without burning up your EL84s? If so, how did you do it?
I'm thinking of fitting a CLC after the UF4007s, then a second RC stage consisting of a 270R 15W resistor and 470uF 450V capacitor. But it sure is wasteful. Seriously wasteful.
hi,
Install a high voltage control, something like that can be quickly soldered together for testing!
https://stromrichter.org/showthread.php?tid=3116&pid=232449#pid232449
or something like that, layout is available
https://www.changpuak.ch/electronics/VB408.php
good luck
Install a high voltage control, something like that can be quickly soldered together for testing!
https://stromrichter.org/showthread.php?tid=3116&pid=232449#pid232449
or something like that, layout is available
https://www.changpuak.ch/electronics/VB408.php
good luck
Thanks everyone for the suggestions and warnings. All good stuff.
And hi again Francois -- As you can see, I'm battling with the parts I have at hand and how to fit them into the chassis I have to work with.
I would like to use transformers I already have, if at all possible. I have the iron from a Dynaco Stereo 35 PP EL84 amp, which I thought would work easily for an EL84 Baby Huey. But you know, things are never as simple as we'd like them to be.
From what I have in stock, my power transformer choices are the PA774 from the Dynaco ST35, or an Edcor XPWR065-120.
The Edcor PT will give me about +325V DC for the raw B+, which is perfect for EL84s. The chassis I have has the mounting holes to fit already drilled. I was hoping to use that transformer for a stereo PP EL86 design, but what the heck, if the shoe fits...
Physically, the PA774 transformer would fit well on this chassis, but there's that nasty problem with the high voltage.
I could add a rectifier tube to knock down the plate voltage, but I'd have to add a 5V AC filament transformer that can deliver 2A, enough for a 5AR4. The octal tube and filament transformer would take up a lot of room in this chassis.
I could also add a decent sized choke and make a choke input supply, or a quasi-choke input supply with a small cap on the input to adjust the output voltage. I have a Hammond 6H choke with 92 ohm DCR, but I really don't know if it will saturate with a steady 160mA going through it. I have no specs for it, as it was pulled from a Traynor YBA-3 Custom Special bass guitar amplifier. The choke fed the screen grid supply for a quad of 7027A output tubes (not the plate supply). I figure the choke is not rated for continuous duty with 160mA plate current going through it.
So, with all this complication, I'm thinking I should follow the path of least resistance and use the Edcor 500VCT 250mA transformer with the Dynaco Z565 OPTs, with good old EL84s.
Maybe I'll use the PA774 transformer for a push-pull 6V6 amp or something, some other time.
Strange that Dynaco chose to use such a high voltage for EL84s. I guess they were cheap back in the day.
And hi again Francois -- As you can see, I'm battling with the parts I have at hand and how to fit them into the chassis I have to work with.
I would like to use transformers I already have, if at all possible. I have the iron from a Dynaco Stereo 35 PP EL84 amp, which I thought would work easily for an EL84 Baby Huey. But you know, things are never as simple as we'd like them to be.
From what I have in stock, my power transformer choices are the PA774 from the Dynaco ST35, or an Edcor XPWR065-120.
- The PA774 has a 660VCT 180mA secondary, and two 6.3V 2.5A filament windings.
- The XPWR065-120 has a 500VCT 250mA secondary, along with a single 6.3V 4A filament winding.
- Neither of these transformers has a 5V winding for a rectifier tube.
The Edcor PT will give me about +325V DC for the raw B+, which is perfect for EL84s. The chassis I have has the mounting holes to fit already drilled. I was hoping to use that transformer for a stereo PP EL86 design, but what the heck, if the shoe fits...
Physically, the PA774 transformer would fit well on this chassis, but there's that nasty problem with the high voltage.
I could add a rectifier tube to knock down the plate voltage, but I'd have to add a 5V AC filament transformer that can deliver 2A, enough for a 5AR4. The octal tube and filament transformer would take up a lot of room in this chassis.
I could also add a decent sized choke and make a choke input supply, or a quasi-choke input supply with a small cap on the input to adjust the output voltage. I have a Hammond 6H choke with 92 ohm DCR, but I really don't know if it will saturate with a steady 160mA going through it. I have no specs for it, as it was pulled from a Traynor YBA-3 Custom Special bass guitar amplifier. The choke fed the screen grid supply for a quad of 7027A output tubes (not the plate supply). I figure the choke is not rated for continuous duty with 160mA plate current going through it.
So, with all this complication, I'm thinking I should follow the path of least resistance and use the Edcor 500VCT 250mA transformer with the Dynaco Z565 OPTs, with good old EL84s.
Maybe I'll use the PA774 transformer for a push-pull 6V6 amp or something, some other time.
Strange that Dynaco chose to use such a high voltage for EL84s. I guess they were cheap back in the day.
Last edited:
Interesting idea! I found a description of using the DC resistance of a power supply choke to derive a negative bias voltage in MerlinB's power supplies book. The questions I have about that are:
1) Will the choke still be passing the full plate current of the four EL84s and two 12AX7s? If so, that's going to be about 160mA at quiescence, and more on signal peaks. If so, the bias voltage at quiescent operation would be 92R(0.16A) = -14.7V. But that voltage will go more negative under peak demands, e.g., -16.6V if the EL84s draw 180mA. It looks like that could be an issue if it causes the cathode voltage of the LTP to vary with changes in signal output from the amplifier.
2) Will that voltage dropped across the choke's DC resistance subtract from the rectified DC voltage? If I put a power resistor in series with the choke to increase the apparent DCR, thus dropping more volts there, will that also decrease the B+ voltage referenced to 0V ground? Or does that not affect the B+ voltage since the first filter capacitor will be referenced to 0V?
________________________________________________________
I found an Aiken Amps article on back biasing that looks promising:
https://www.aikenamps.com/index.php/what-is-back-biasing
The following looks like it could be a good way to derive a -20V bias supply for the tail of the 12AX7 LTP (I'd use a 24V Zener instead of the 15V Zener pictured):
However, if I increase the value of C2 way up to 100uF or more, would the inrush current at turn on burn out D3?
Hmmm, there might be a fly in the ointment/ The B+ supply will be drawing 160mA or so at quiescence, and more during peak demand. If 24V is dropped across the zener with 0.16A, that's 3.84 watts of dissipation. Now I'm looking at a zener string that may need a heat sink. Complications...
______________________________________________________
I was just going to go with a brute force solution of deriving a B- supply using silicon diodes with a couple of decoupling stages to knock the voltage down to something reasonable, then using a large value resistor as the tail load. Like this (270k 2W in the schematic below):
That's basically what I'm planning to build using the Edcor 500VCT power transformer, but the B+ and B- voltages may be more like +/- 310V after filtering stages. The downside is the need for the very large value resistor in the LTP tail, which will exceed the 12AX7 cathode to heater voltage until the tubes warm up. That's why the 1N4148 diode is connected from 12AX7 cathodes to 0V. I'm pretty sure that will work well enough.
--
1) Will the choke still be passing the full plate current of the four EL84s and two 12AX7s? If so, that's going to be about 160mA at quiescence, and more on signal peaks. If so, the bias voltage at quiescent operation would be 92R(0.16A) = -14.7V. But that voltage will go more negative under peak demands, e.g., -16.6V if the EL84s draw 180mA. It looks like that could be an issue if it causes the cathode voltage of the LTP to vary with changes in signal output from the amplifier.
2) Will that voltage dropped across the choke's DC resistance subtract from the rectified DC voltage? If I put a power resistor in series with the choke to increase the apparent DCR, thus dropping more volts there, will that also decrease the B+ voltage referenced to 0V ground? Or does that not affect the B+ voltage since the first filter capacitor will be referenced to 0V?
________________________________________________________
I found an Aiken Amps article on back biasing that looks promising:
https://www.aikenamps.com/index.php/what-is-back-biasing
The following looks like it could be a good way to derive a -20V bias supply for the tail of the 12AX7 LTP (I'd use a 24V Zener instead of the 15V Zener pictured):
However, if I increase the value of C2 way up to 100uF or more, would the inrush current at turn on burn out D3?
Hmmm, there might be a fly in the ointment/ The B+ supply will be drawing 160mA or so at quiescence, and more during peak demand. If 24V is dropped across the zener with 0.16A, that's 3.84 watts of dissipation. Now I'm looking at a zener string that may need a heat sink. Complications...
______________________________________________________
I was just going to go with a brute force solution of deriving a B- supply using silicon diodes with a couple of decoupling stages to knock the voltage down to something reasonable, then using a large value resistor as the tail load. Like this (270k 2W in the schematic below):
That's basically what I'm planning to build using the Edcor 500VCT power transformer, but the B+ and B- voltages may be more like +/- 310V after filtering stages. The downside is the need for the very large value resistor in the LTP tail, which will exceed the 12AX7 cathode to heater voltage until the tubes warm up. That's why the 1N4148 diode is connected from 12AX7 cathodes to 0V. I'm pretty sure that will work well enough.
--
Wonder why @rongon wants to build a simplified Baby Huey? Similar to the first iteration @gingertube stated out with, before adding on all the “improvements”. Let’s see. 
Wonder if simpler is better?
You have the Baby Huey PC’s (I thought), you have great Dynaco Z565 UL output transformers, you have a suitable 500Vac CT power transformer. All you need to get is a small 30-50 Vac transformer (~$15) for the bias supply and you can build you very own “full-fledged” Baby Huey.
Wonder if simpler is better?You have the Baby Huey PC’s (I thought), you have great Dynaco Z565 UL output transformers, you have a suitable 500Vac CT power transformer. All you need to get is a small 30-50 Vac transformer (~$15) for the bias supply and you can build you very own “full-fledged” Baby Huey.
It's an experiment.
Also, the PCBs won't fit in this chassis without surgery, since this chassis has already been punched and drilled for six 9-pini mini tube sockets, with mounting holes and space for two OPTs and a single power transformer.
I also like the idea that this simplified version has fewer active parts, so should be easier to troubleshoot should problems arise.
For the Baby Huey PCBs, I'd like to start from scratch, punching the holes in the right places the first time.
Also, the PCBs won't fit in this chassis without surgery, since this chassis has already been punched and drilled for six 9-pini mini tube sockets, with mounting holes and space for two OPTs and a single power transformer.
I also like the idea that this simplified version has fewer active parts, so should be easier to troubleshoot should problems arise.
For the Baby Huey PCBs, I'd like to start from scratch, punching the holes in the right places the first time.
rongon,
Experiments are good to do.
And so is reasonable simplicity.
1. Three feedback loops are probably enough, or more than enough:
Ultra Linear (local)
Schade (nearly local)
Output transformer secondary to phase inverter grid (global).
2. There is a very good reason for the -320V supply, it is required by the 270k LTP resistor.
Using a negative supply of only -20V is going to need a solid state CCS.
From -20V, any LTP resistor will be too low of a resistance (16.8k). With only 16.8k, the cathode signal current of the driven triode will be 'robbed' by the "LTP" resistor, instead of going to the non driven triode's cathode of the phase inverter (the global negetive feedback signal does not properly fix that problem).
If you can not provide a -320V supply, then use an LM334 which will work very well with a -20V supply and the 1.2mA required by the 12AX7 pair of triodes.
Note: even with a -20V supply, you still need a 1N4148 there to protect the 12AX7s grids from a Hot Start.
3. You might try looking at a square wave on a scope, and then trying a small capacitor across the 12k to 20k global negative feedback resistor.
Should be fun!
Experiments are good to do.
And so is reasonable simplicity.
1. Three feedback loops are probably enough, or more than enough:
Ultra Linear (local)
Schade (nearly local)
Output transformer secondary to phase inverter grid (global).
2. There is a very good reason for the -320V supply, it is required by the 270k LTP resistor.
Using a negative supply of only -20V is going to need a solid state CCS.
From -20V, any LTP resistor will be too low of a resistance (16.8k). With only 16.8k, the cathode signal current of the driven triode will be 'robbed' by the "LTP" resistor, instead of going to the non driven triode's cathode of the phase inverter (the global negetive feedback signal does not properly fix that problem).
If you can not provide a -320V supply, then use an LM334 which will work very well with a -20V supply and the 1.2mA required by the 12AX7 pair of triodes.
Note: even with a -20V supply, you still need a 1N4148 there to protect the 12AX7s grids from a Hot Start.
3. You might try looking at a square wave on a scope, and then trying a small capacitor across the 12k to 20k global negative feedback resistor.
Should be fun!
Last edited:
Layout decisions first. Here's what the transformers look like on this hand-me-down chassis I have:
Is the orientation of the power transformer OK in relation to the output tubes?
Does that 6H choke look capable of handling two channels of push-pull EL84s?
I'll have to drill some new holes for the OPTs and the LTP tube sockets, and it looks like I'll have to punch holes to fit a couple of large 800uF 330V radial capacitors I have in stock.
I'm thinking the power supply will consist of a simple full-wave rectifier (power transformer CT secondary) into a 100uF 450V reservoir cap, a choke, and a 470uF 450V filter cap. A bleeder resistor would go there too. From there, each channel gets its own 150 ohm 5W resistor and 800uF 330V capacitor decoupling for its plate supply (B+), I'm pretty sure that will be enough ripple filtering.
I have two chokes I can use for this. The 6H 92R one pictured above, or a Hammond 159T 'reactor' (swinging choke), rated for 2.5H at 300mA. I know that one will work. Here's a photo of the 6H mystery choke and the Hammond 159T, side by side:
Going by size alone:
Can one infer the continuous current capacity of the chokes from their core size?
If so, I'd guess the 6H choke is rated for about 150mA continuous, and so is just marginal for use in this circuit. Or am I wrong about that?
For the LTP tail supply, it's actually easier to provide a -320V (or so) DC supply than it is to provide a -30V DC supply.
The -320V DC supply (B-) will come from another pair of silicon rectifier diodes with a CRC filter network, then each channel's 12AX7 gets a 270k 2W resistor to its 'tail'.
I have a 12VCT 0.45A filament transformers I could use to make an approximately -16V DC supply for a pair of LM334 current sinks. It's small, so I know I could fit in there somewhere. The downside would be a second 120V primary to wire up in this small chassis, which could inject hum into the audio circuitry. It just seems a safer bet to run -320V DC to the 12AX7s.
That's what I have so far...
Is the orientation of the power transformer OK in relation to the output tubes?
Does that 6H choke look capable of handling two channels of push-pull EL84s?
I'll have to drill some new holes for the OPTs and the LTP tube sockets, and it looks like I'll have to punch holes to fit a couple of large 800uF 330V radial capacitors I have in stock.
I'm thinking the power supply will consist of a simple full-wave rectifier (power transformer CT secondary) into a 100uF 450V reservoir cap, a choke, and a 470uF 450V filter cap. A bleeder resistor would go there too. From there, each channel gets its own 150 ohm 5W resistor and 800uF 330V capacitor decoupling for its plate supply (B+), I'm pretty sure that will be enough ripple filtering.
I have two chokes I can use for this. The 6H 92R one pictured above, or a Hammond 159T 'reactor' (swinging choke), rated for 2.5H at 300mA. I know that one will work. Here's a photo of the 6H mystery choke and the Hammond 159T, side by side:
Going by size alone:
- Both have the same size core at 3.5" x 2".
- the 6H choke's laminations stack is ever-so-slightly fatter at 1-1/16" thick.
- The Hammond 159T (2.5H 300mA) choke has a slightly thinner lam stack at 1" thick.
Can one infer the continuous current capacity of the chokes from their core size?
If so, I'd guess the 6H choke is rated for about 150mA continuous, and so is just marginal for use in this circuit. Or am I wrong about that?
For the LTP tail supply, it's actually easier to provide a -320V (or so) DC supply than it is to provide a -30V DC supply.
The -320V DC supply (B-) will come from another pair of silicon rectifier diodes with a CRC filter network, then each channel's 12AX7 gets a 270k 2W resistor to its 'tail'.
I have a 12VCT 0.45A filament transformers I could use to make an approximately -16V DC supply for a pair of LM334 current sinks. It's small, so I know I could fit in there somewhere. The downside would be a second 120V primary to wire up in this small chassis, which could inject hum into the audio circuitry. It just seems a safer bet to run -320V DC to the 12AX7s.
That's what I have so far...