This is the start of a thread for a P-P amp using the power and output iron from a Fisher 500B receiver. I got the transformers a couple of years ago on Ebay, intending to do a 7591A amp with radically different input/driver stage than the original design, which used 12AX7 for input and concertina phase splitter.
In short, I'm using a hybrid differential input stage featuring a pair of matched J309 JFETs cascoded with the small triode each from a pair of 6CS7 dual dissimilar triodes. The rationale behind this is to get some decent first stage gain and allow easy close matching of the input differential piar - a snap fro me with JFETs, not quite so easy (or cheap) with tubes. I'll be using a similar scheme with a pair of JFETs and a 12AU7A to drive a pair of 6P14P-EVs for another project.
So that the second, beefier triode does not feel left out, the remaining two triodes are used as source folowers to drive the 7591A control grids. This will allow some excursion into Class AB2 operation without blocking and other antisocial behavior.
Actually the biggest pain so far has been to put together the dropping strings to generate all the fiddly intermediate voltages for the 7591A screens, a cleaner supply for the input stage, and a ~200V supply for the cathode followers. I could have cheated and built a dedicated switching supply with all those outputs available, but I was determined to use the original power transformer, which has a single HV winding that drove a doubler in the original receiver schematic, generating a +400V supply. In the original receiver, all the B+ voltages were derived from this supply. Same for the new amp.
The circuit and picture will follow when I'm happy with everything. Hopefully the amp will make an appearance at the 2009 Burning Amp this weekend.
In short, I'm using a hybrid differential input stage featuring a pair of matched J309 JFETs cascoded with the small triode each from a pair of 6CS7 dual dissimilar triodes. The rationale behind this is to get some decent first stage gain and allow easy close matching of the input differential piar - a snap fro me with JFETs, not quite so easy (or cheap) with tubes. I'll be using a similar scheme with a pair of JFETs and a 12AU7A to drive a pair of 6P14P-EVs for another project.
So that the second, beefier triode does not feel left out, the remaining two triodes are used as source folowers to drive the 7591A control grids. This will allow some excursion into Class AB2 operation without blocking and other antisocial behavior.
Actually the biggest pain so far has been to put together the dropping strings to generate all the fiddly intermediate voltages for the 7591A screens, a cleaner supply for the input stage, and a ~200V supply for the cathode followers. I could have cheated and built a dedicated switching supply with all those outputs available, but I was determined to use the original power transformer, which has a single HV winding that drove a doubler in the original receiver schematic, generating a +400V supply. In the original receiver, all the B+ voltages were derived from this supply. Same for the new amp.
The circuit and picture will follow when I'm happy with everything. Hopefully the amp will make an appearance at the 2009 Burning Amp this weekend.
Be careful with that Fisher power transformer. My X-101-B draws 208mA quiescent. The Sam's folder shows the power transformer as rated at 200mA.
One day my wife said, "What's that smell?" I ran in to check and noticed wisps of white smoke coming out of that transformer, amp still playing music. I killed the power and listened to the insulation inside the transformer boil for the next few minutes. I guess it never shorted because the fuse never blew. I'll be building an outboard supply as I don't care to have a transformer rewound that was inadequate to begin with. Plus, I've recently fallen in love with Maida regulators for P-P AB power stages.
One day my wife said, "What's that smell?" I ran in to check and noticed wisps of white smoke coming out of that transformer, amp still playing music. I killed the power and listened to the insulation inside the transformer boil for the next few minutes. I guess it never shorted because the fuse never blew. I'll be building an outboard supply as I don't care to have a transformer rewound that was inadequate to begin with. Plus, I've recently fallen in love with Maida regulators for P-P AB power stages.
If it comes to that, I'll gin up a switcher that will run as cool as a cucumber and allow me to strip out all the lousy dropping resistors. I plan to use standard 7591A biasing to start with, 30-35 mA per tube, 400V plate, 350V screen.
That's what I'm hoping, too. Right now, I have a matched quad of the EH 7591As. I had an old Scott LK-72B years ago (never should have sold it). The 7591As in that beast glowed rather fetchingly as well. I forget whether they were Scott branded or Sylvanias. The glow would comoplement my chassis nicely - metallic cast-iron gray for outer chassis and transformers, deep metallic blue-green for the front anf top deck, with blue-green LED in front.
IF the EH tubes poop out on me, I'm pretty sure I have enough latitude in bias and drive to make do with some 6P3S-Es.
IF the EH tubes poop out on me, I'm pretty sure I have enough latitude in bias and drive to make do with some 6P3S-Es.
I found the fisher amp in my mother-in-law's attic. All of the 7591s in the amp had the blue glow. I think two were GE with a regular base, one was a GE with a coin base, and one was marked 'Fisher'. (maybe one of the original tubes?) Unfortunately, the GE coin base 7591 would glow red as well after about 60 seconds on.
The tubes in the picture are new EH 7591s. In my limited experience, I've actually never seen a 7591 that didn't have a nice blue glow.
The tubes in the picture are new EH 7591s. In my limited experience, I've actually never seen a 7591 that didn't have a nice blue glow.
Avoiding blocking distortion is fine. Serious g1 current in 7591s is not a good idea. The type is high gm, which means a tight g1 helix. If substantial g1 current is present, the heat generated has no place to go. 🙁
Thats why I'm using just cathode followers (and not particularly rugged ones) to drive the outputs rather than a mosfet. I wasn't planning on hammering the snot out of the poor 7591s, but to be able to take light dips into AB2 without serious audible consequences.Actual depth of AB2 operation depends on the drive stage, and I'll be taking steps to limit that. I may also consider putting more stopper resistance on the grids to help limit grid current. Right now I have 120 ohms.
Edit -I have the drive capability, so if I feel like hammering, I can always drop some other 6L6 variant in the output sockets and rebias. However, since I have 35W transformers and only 400V B+, there won't be a whole lot of payback from getting too frisky.
Edit -I have the drive capability, so if I feel like hammering, I can always drop some other 6L6 variant in the output sockets and rebias. However, since I have 35W transformers and only 400V B+, there won't be a whole lot of payback from getting too frisky.
Last edited:
I just picked up a X-101-B so I'll be watching this thread with great interest.
I haven't even powered it up yet.
I haven't even powered it up yet.
Wrenchone,
A MOSFET follower is fine. Just avoid the big honking types, like the IRFBC20. The little ZVN0545A should be "perfect", for this job. The triodes you orginally planned on using might go to work as pass devices in series regulators.
BTW, what's your opinion of 10 KOhm 7591 g1 stoppers? That should provide protection against overheating, while still allowing a bit of AB2 action to occur.
The amp is all wired up except for output tubes, and I just finished getting the bias right on the input/driver stage last night (including those cathode followers), so I'm going to go with what I've got for the time being -especially as I'm trying to get the amp finished for Burning Amp this weekend.
The cathode followers were a case of "the tube is there, so use it". If I were courting serious screen current, as in a screen drive P-P amp (in the works later on), I would have planned on mosfet follwers from the beginning.
I'm putting 22V clamp zeners across my input JFETS to protect them in cases of extreme drive situations where the feedback unhooks briefly (especially as the fets I'm using have only 25V VDS rating). This also serves to limit the excursion of the input differential stage under extreme conditions, preventing the output tubes from being hammered too hard.
I'll have to think about using 10k for stoppers...
For those seeking to do something similar with 7591As and spare iron - well, this circuit is fairly complex (you'll see later), and an easy-to drive-tube like the 7591A could well be served with a less complex circuit.
I can think of three off the top of my head. The first would be a simpler version of the circuit I'm trying here, omitting the cathode folower. This would leave a jfet differential stage cascoded with a double triode (a 12AU7 would work nicely). Just three tubes per channel, and a lot less fol-de roll with dropping resistors and bypass caps - simpler biasing, to boot.
Second would be a pentode input stage with a triode concertina splitter using something like the 6AW8).
A third would be a triode input stage with concertina splitter using a tube like the 6JK8, which has a nice high transconductance low plate resistance triode (input duty) and a more modest second triode that would make an ok phase splitter.
Lot of things would work. the original Fisher circuit used a 12AX7 for input and phase splitter and managed to sound OK.
The cathode followers were a case of "the tube is there, so use it". If I were courting serious screen current, as in a screen drive P-P amp (in the works later on), I would have planned on mosfet follwers from the beginning.
I'm putting 22V clamp zeners across my input JFETS to protect them in cases of extreme drive situations where the feedback unhooks briefly (especially as the fets I'm using have only 25V VDS rating). This also serves to limit the excursion of the input differential stage under extreme conditions, preventing the output tubes from being hammered too hard.
I'll have to think about using 10k for stoppers...
For those seeking to do something similar with 7591As and spare iron - well, this circuit is fairly complex (you'll see later), and an easy-to drive-tube like the 7591A could well be served with a less complex circuit.
I can think of three off the top of my head. The first would be a simpler version of the circuit I'm trying here, omitting the cathode folower. This would leave a jfet differential stage cascoded with a double triode (a 12AU7 would work nicely). Just three tubes per channel, and a lot less fol-de roll with dropping resistors and bypass caps - simpler biasing, to boot.
Second would be a pentode input stage with a triode concertina splitter using something like the 6AW8).
A third would be a triode input stage with concertina splitter using a tube like the 6JK8, which has a nice high transconductance low plate resistance triode (input duty) and a more modest second triode that would make an ok phase splitter.
Lot of things would work. the original Fisher circuit used a 12AX7 for input and phase splitter and managed to sound OK.
I have one side of the amp basically up after fixing some bonehead mistakes. Square wave response looks good, but the amp has a tendency to burst into oscillation at inopportune moments. When it's not going full tilt, there's a nasty snarky burst of oscillation at the output every AC 1/2 cycle. Oscillation is about 2 MHz. I traced it to the plate of one of my 7591As (may be at the other plate as well - I didn't look. Unlike other circuits I've tried, the JFET input stage looks well-behaved. I suspect inadequate grid stopper values, as mine are 120 ohms and the Fisher schematic uses 1k, probably with good reason.
There are enough annoyances so that I'm probably going to shelve this amp until after Burning Amp, so that I can concentrate on doing a little cleaning up on projects that are working for sure. I also don't like the internal layout all that much,so I might do a re-layout on the circuit boards I'm using in the amp and split out some functions so I can place things better and cut down the length of some interconnects..
Anyway, first on the list when I get back to this amp will be to try some respectable grid stoppers of 1-2k. Would the tube benefit from some screen stoppers as well? Suggested values? How about a plate-to-plate RC damper? This is my first foray into P-P amps with high transconductance pentodes to boot, so suggestions are welcome.
There are enough annoyances so that I'm probably going to shelve this amp until after Burning Amp, so that I can concentrate on doing a little cleaning up on projects that are working for sure. I also don't like the internal layout all that much,so I might do a re-layout on the circuit boards I'm using in the amp and split out some functions so I can place things better and cut down the length of some interconnects..
Anyway, first on the list when I get back to this amp will be to try some respectable grid stoppers of 1-2k. Would the tube benefit from some screen stoppers as well? Suggested values? How about a plate-to-plate RC damper? This is my first foray into P-P amps with high transconductance pentodes to boot, so suggestions are welcome.
Amp project not dead, just sleeping on a piece of foam so as not to mar the paint. Still managed a bo-bo, which may be fixable with some car wax.
Anyway, even before Burning Amp, I managed to re-do the layout on my driver boards, splitting the input differential functions and cathode follower current source loads into two separate smaller boards, which should make it much easier to shorten up all the interconnects. This, and changing the output grid stoppers (and adding screen stoppers) should make the amp much easier to manage. I wished I had the thing at BA this year - 30WPC would have gone some way towards cutting through the ambient noise and the sound-eating quality of the big rooms there, as well as dealing with some small but power hungry speakers.
Spread Spectrum - on the one side I fully populated and tested, I was getting some nice blue glow on the 7591As before I closed up shop to concentrate on lower hanging fruit for BA. I'm looking froward to having the amp right side up, buttoned down, and playing.
Anyway, even before Burning Amp, I managed to re-do the layout on my driver boards, splitting the input differential functions and cathode follower current source loads into two separate smaller boards, which should make it much easier to shorten up all the interconnects. This, and changing the output grid stoppers (and adding screen stoppers) should make the amp much easier to manage. I wished I had the thing at BA this year - 30WPC would have gone some way towards cutting through the ambient noise and the sound-eating quality of the big rooms there, as well as dealing with some small but power hungry speakers.
Spread Spectrum - on the one side I fully populated and tested, I was getting some nice blue glow on the 7591As before I closed up shop to concentrate on lower hanging fruit for BA. I'm looking froward to having the amp right side up, buttoned down, and playing.
And I was hoping for beer. That's all I drink in India when I'm there since you can't drink the water.
I have another set of 500B outputs and a complete set of 400 iron. An obvious next pairing would be some 6P3S-Es with some of this iron...
A friend gave me a Fisher receiver (500B ?) that he "found in his attic". Now attics in Florida are often in the 130 to 150 degree range during the daylight hours, and this RX had been there for at least 10 years, so I didn't expect much. I planned to relieve it of its OPT's and 7591's but I decided to just plug it in and see what happened. It works! So it would now be a crime to dismantle it, and I don't need it so it hangs out in the warehouse with a few other pieces that I have collected over the years.
I have never done much experimentation with 7591's. I can tell you that ANY grid stopper at all will cause measurable and audible distortion when the grid goes positive. The input impedance of G1 is megohms in parallel with C miller until you start to get into the -1 to -3 volt region. Here it gently drops with a logarithmic slope until you start to go positive. The grid impedance in the positive voltage area abruptly drops (depending on the tube) to a low and constantly decreasing value. I have measured grid impedances in the hundreds of ohms on an 833A.
My recent AB2 experiments with 6L6GC's don't indicate this extreme value, but I will measure it when I get a chance. I have no grid stoppers in my latest experiments, just a few ferrite beads on the grid wires.
I'm trying 1k resistors right now as stoppers on the Kingfisher, as well as 120 ohm screen stoppers. Both are deployed right at the socket. The way I have the grids stoppers installed, it would be a simple matter to try beads later on. Some sort of stopper does appear to be necessary to avoid unpredictible full-tilt bursts of oscillation. Fisher used 1k in their 7591A designs. I'm not going for full-tilt AB2 operation.
As far as that 500B receiver is concerned, I would refurbish it a little at a time and find somebody competent to align the tuner - they are too nice to gut. I just bought my iron off Ebay (bargain prices, too), The 400 transformer set I picked up at a local surplus store (I came, I saw, I grabbed).
As far as that 500B receiver is concerned, I would refurbish it a little at a time and find somebody competent to align the tuner - they are too nice to gut. I just bought my iron off Ebay (bargain prices, too), The 400 transformer set I picked up at a local surplus store (I came, I saw, I grabbed).
- Status
- Not open for further replies.