Ok i got a guy at work who wants to rebuild a gutted fisher stereo el84pp amp. All it has left is the sockets and transformers. so each side has 3 sockets, 2 for the el84's and one for a 9pin 12ax7 or something. The output transformers are 8k and the b+ is 320v unloaded.
He wants to modernize it with low gain(running off an older nad cd player) and run it in pentode mode for 15-18 watts to run some 85db book shelfs.
I have all the parts for a diyparadise simple el84 pp with a lm317 in the tail for the power stage, but i got nothing for the driver stage, and i dont know how to enable a feedback loop, which i know he will need since he will need since he wants pentode mode.
Any help with a simple driver stage with feedback (do they made local feedback output stages)? I think he would be happy with a el84 for the driver since it would mean the same tube for everything.
He wants to modernize it with low gain(running off an older nad cd player) and run it in pentode mode for 15-18 watts to run some 85db book shelfs.
I have all the parts for a diyparadise simple el84 pp with a lm317 in the tail for the power stage, but i got nothing for the driver stage, and i dont know how to enable a feedback loop, which i know he will need since he will need since he wants pentode mode.
Any help with a simple driver stage with feedback (do they made local feedback output stages)? I think he would be happy with a el84 for the driver since it would mean the same tube for everything.
Putting a LM317 in the O/P stage tail forces Class "A" operation. So, expect only 10 WPC. A yield of 15 WPC is about right from Class "AB" pentode mode. Either way, background music is what he'll get with those 85 dB. speakers. Per Paul Joppa's 102 dB. rule, 85 dB. speakers need at least 60 WPC, for serious listening.
Assuming the limited power yield is OK, there's more than 1 way to skin this cat. Naturally, I'm partial to things I've been associated with. An "El Cheapo" variant, with "fixed" bias pentode mode O/P tubes, should be satisfactory. The 12AT7 splitter/driver will require ZVN0545A source follower buffering. By any chance, does the Fisher power trafo have a bias winding or tap?
Assuming the limited power yield is OK, there's more than 1 way to skin this cat. Naturally, I'm partial to things I've been associated with. An "El Cheapo" variant, with "fixed" bias pentode mode O/P tubes, should be satisfactory. The 12AT7 splitter/driver will require ZVN0545A source follower buffering. By any chance, does the Fisher power trafo have a bias winding or tap?
Attachments
Hmm i have a lm317 setup from diyaudioprojects that uses a 300v b+ instead of 210 b+ in pentode mode. i just assumed it was putting out 15 watts instead of 7watts due to the higher b+ and pentode move vs stock ul mode. Oh well.
http://diyparadise.com/buildel84b.html
Thats not 15 watt in pentode mode? Again he doesnt have a feed back loop, but he uses ul or triode mode so its not needed as much.
Yes the power transformer has a 2 wire bias out that is 40v.
http://diyparadise.com/buildel84b.html
Thats not 15 watt in pentode mode? Again he doesnt have a feed back loop, but he uses ul or triode mode so its not needed as much.
Yes the power transformer has a 2 wire bias out that is 40v.
Last edited:
Repeat, the tail CCS forces Class "A" operation. Theory tells us that 50% efficiency is max. for Class "A". Pentodes are pretty darned good real world and will deliver 40% efficiency, in Class "A". If each tube's plate is dissipating the max. allowable 12.5 W., the amp delivers approx. 10 W. To get 15 WPC, the net efficiency must be 60%. Theory states 78.5% is max. in Class "B". A Class "AB" operating point, which is necessary for decent sound, must deliver lower efficiency. Getting 15 WPC and decent sound from a PP pair of EL84s requires considerable care.
The separate 40 V. winding is GOOD news. That winding can be bridge rectified to provide B- ('T7 tail CCS) and C- (EL84 control grids).
BTW, full pentode mode "finals" work best in combination with regulated g2 B+. A Maida regulator will take care of that item, quite nicely.
How did Fisher rectify the B+: FWCT, bridge, or Greinacher doubler? Was the "donor" an integrated amp, with a phono section and tone controls?
A common CCS -- and i was told only works in triode (Allen Wright). A pair, as in Baby Huey, allows AB operation.
dave
LOL i dont want to start another argument thread. I just wanted to get a simple amp going. If its going to take more wiring and parts to get another 5 watts its not worth it. thats not even 2db.
Still on topic of a pentode setup with some feed back so he doesnt get too sloppy of output.
The b+ is FWCT, i was planning on converting it to SS Since the whole unit is tubeless right now. No the amp has no controls, just 2 rca in's.
Still on topic of a pentode setup with some feed back so he doesnt get too sloppy of output.
The b+ is FWCT, i was planning on converting it to SS Since the whole unit is tubeless right now. No the amp has no controls, just 2 rca in's.
Snubbed UF4007s will do quite well to rectify the B+. Keep the 1st filter cap. size reasonably small, to prevent trafo overheating, and install a Triad C-24X choke between the 1st cap. and the reservoir part.
Why bother with a CCS in the O/P stage tail, to begin with? Believe me, between infrasonic filtering at the amp's I/Ps and the GNFB circuitry, you'll do just fine. Remember, Fisher built with good "iron".
Why bother with a CCS in the O/P stage tail, to begin with? Believe me, between infrasonic filtering at the amp's I/Ps and the GNFB circuitry, you'll do just fine. Remember, Fisher built with good "iron".
El cheapo looks good. I got the negative bias supply, But can the ccs be replaced my something simpler?
YH's Lazy Man Constant Current Source
LM317 / LM338 / LM350 Voltage and Current Regulator Calculators
And do i need a cap at the input?
And what if i dont need a volume pot, or just have left over 50k's or something. I can just omit it and not change any other values right?
YH's Lazy Man Constant Current Source
LM317 / LM338 / LM350 Voltage and Current Regulator Calculators
And do i need a cap at the input?
And what if i dont need a volume pot, or just have left over 50k's or something. I can just omit it and not change any other values right?
Why not keep it simple and use something like a 12AT7 or 6N1P as a CC VAS and cathodyne PI with feedback from OPT secondary to VAS cathode? The only really tricky part I see is getting a stable amp with gNFB without a scope but you are going to need some kind of FB with pentode mode. The other option would be to use a LTP PI and plate to plate FB over the EL84s which won't have as much stability concern as global. For LTP you could go CCS as mentioned or add a small PT and some Si diodes for a negative rail and use a big resistor.
I think that you would have to abuse the EL84s a bit to get 18 hi-fi watts out of it but a dozen should be doable and adequate for a reasonable size room.
I think that you would have to abuse the EL84s a bit to get 18 hi-fi watts out of it but a dozen should be doable and adequate for a reasonable size room.
Ok that first sentence was waaaaay over my head.
This is only going to be full tube amp #4 for me, and #3 is still in production and waiting on edcor (COME ON I ORDERED MID FEB!!!)
This is only going to be full tube amp #4 for me, and #3 is still in production and waiting on edcor (COME ON I ORDERED MID FEB!!!)
You can use a 10M45S integrated circuit as the constant current sink (CCS).
Yes, you do. The cap. combines with the grid to ground resistor to keep infrasonic garbage out of the amp. That, in turn, protects against O/P trafo core saturation, due to an overly large GNFB LF error correction signal. A capacitance of 0.068 μF. combines with 100 KOhms to yield a F3 of 23.4 Hz. That should be excellent in combination with the targeted bookshelf speakers. BTW, a 25 WVDC part is more than tough enough here.
Only if you are building an "integrated" amp is the volume control necessary.
Sorry. Fred's PPP shown here uses some of what I was talking about and will do for a quick discussion.
http://www.dogstar.dantimax.dk/tubestuf/graphics/hk-retro.gif
The 12AX7 in his schematic is the CC (Common Cathode) VAS (Voltage Amplification Stage) that I mentioned. Notice that he feeds back some of the output voltage from the output transformer secondary through R118 to the cathode of the 12AX7 (R116). Those two resistors make up a voltage divider that applies some of the output to the cathode of the input tube. That is the gNFB (global Negative Feedback) loop.
The 12AU7 in his diagram is a cathodyne PI (Phase Inverter). He biases it with the voltage divider (R113 and R114) but there are other ways such as carefully selecting the bias of the first stage and directly coupling to the grid of the PI among others.
I was suggesting using a single 6N1P or 12AT7 for both of these functions.
Of course his output stage is way different (sorry best example I could find in a quick search) as you will be using a single pair of EL84s instead of six 35C5s and will be running class AB with bypassed cathode resistors for bias. Personally I would consider using a separate cathode resistor and bypass cap for each EL84.
As to your question of whether the CCS can be replaced with something simpler... yes. If you use a larger negative supply you can just replace it with a large resistor. The larger the resistor the closer you come to a CCS.
http://www.dogstar.dantimax.dk/tubestuf/graphics/hk-retro.gif
The 12AX7 in his schematic is the CC (Common Cathode) VAS (Voltage Amplification Stage) that I mentioned. Notice that he feeds back some of the output voltage from the output transformer secondary through R118 to the cathode of the 12AX7 (R116). Those two resistors make up a voltage divider that applies some of the output to the cathode of the input tube. That is the gNFB (global Negative Feedback) loop.
The 12AU7 in his diagram is a cathodyne PI (Phase Inverter). He biases it with the voltage divider (R113 and R114) but there are other ways such as carefully selecting the bias of the first stage and directly coupling to the grid of the PI among others.
I was suggesting using a single 6N1P or 12AT7 for both of these functions.
Of course his output stage is way different (sorry best example I could find in a quick search) as you will be using a single pair of EL84s instead of six 35C5s and will be running class AB with bypassed cathode resistors for bias. Personally I would consider using a separate cathode resistor and bypass cap for each EL84.
As to your question of whether the CCS can be replaced with something simpler... yes. If you use a larger negative supply you can just replace it with a large resistor. The larger the resistor the closer you come to a CCS.
Ok last question of the day for elcheapo. Whats the current im shooting for, for the ccs, and if i replace it with the 10m45s, do i need to keep the 10k resistor in the tail too?
Do you want the preamp part? I think all Fishers don't have UL on their Output transformers. If you want something simple and just an amp look the Eico HF-86 amp. Eico HF-81 with preamp is also well liked if that is what you need.
Heres one I've just built. Falls below you power requirements, but is an easy build and sounds great:
unexpectedly good el84 amp
unexpectedly good el84 amp
Set whatever you use for the tail CCS to operate at 6 mA.
That 10 KOhm Carbon film (CF) part in the LTP tail does 2 jobs: it thermally isolates the SS stuff from the tube and it drops unwanted Volts. As your negative supply will be "shorter" than that shown in the graphic, use a 100 Ω CF part in that position for thermal isolation.
--------------------------------------------------------------------------------------------------------------------
Construct the B+ supply and load it down with resistance. If the loaded rail comes in at roughly 355 V., you can use cathode bias, as shown in the graphic. Should the rail be substantially smaller, "fixed" bias EL84s will be necessary. Along with "fixed" bias, the LTP load resistors have to come down to 33 KOhms and ZVN0545A source follower buffering is required. The 'T7 sounds GOOD with a 3 mA. IB and 200 to 220 V. on the plate. Buffering is needed when the load resistor is only 33 KOhms to keep gain up and distortion down. Another reason for the source followers, when "fixed" bias is used, is the reduced control grid to ground resistor limit. Refer to the EL84/6BQ5 data sheet. When the g1 to ground resistance is reduced, the interstage coupling caps. have to increase in value. Using 0.33 μF. coupling caps. will allow the total g1 to ground resistance to go as low as 100 KOhms.
Another +1 for the one Woodturner Fran mentioned
This one is pretty much ready to go out of the box. I implemented a couple tweaks mentioned on diyparadise regarding adding adjustment pots and putting a CCS on the el84 cathodes. I have also replaced the 340 ohm first section 6n1p cathode resistors with red LEDs.
I have tried 6n1p, 6922, 6dj8, and 6cg7 in the driver stage - so far the 6cg7s are my favorite.
Also noteworthy is the 400 ohm el84 cathode resistor is a mistake...it should be 200 ohm if you were to build it as published from JB...
This one is pretty much ready to go out of the box. I implemented a couple tweaks mentioned on diyparadise regarding adding adjustment pots and putting a CCS on the el84 cathodes. I have also replaced the 340 ohm first section 6n1p cathode resistors with red LEDs.
I have tried 6n1p, 6922, 6dj8, and 6cg7 in the driver stage - so far the 6cg7s are my favorite.
Also noteworthy is the 400 ohm el84 cathode resistor is a mistake...it should be 200 ohm if you were to build it as published from JB...
Last edited:
- Status
- Not open for further replies.