I bought a ST70 off fleabay that was original and have done a few mods to it. The first thing I did was replace the selenium rectifier with a diode , I didnt have to add or change any resistors to get the bias to adjust to the proper voltage. I also added two diodes from the power transformer to the 5AR4 tube. I replaced the choke and multicap. I went with a larger 80,40,30,20 version. I also placed a 100 ohm 1/4 watt resistor between pin 4 (screen ?) and the output transformer on each of the output tubes. On the drive board the only thing I changed were the coupling caps from the original (1960) to modern russian k40 PIO caps, I kept the values the same as original. I also added a .047 mfd cap to each input to block any DC input. Is there anything else that should be done, the amp is driving a pair of Magnepan 1.6 speakers ands sounds better than my Yamaha A-1020.
Replace the screen grid resistors with 2Watt as the EL34 can draw screen current when driven hard!
A good basic Ultra Linear Amp.
A good basic Ultra Linear Amp.
The ST-70 does not use screen grid resistors, and hasn't really demonstrated a need to have them.
http://www.dynakitparts.com/image/data/Dyna-ST70.pdf (This is the original version of the assembly manual)
http://www.dynakitparts.com/image/data/Dyna-ST70.pdf (This is the original version of the assembly manual)
The screen resistors are not part of the original design and were added in case of tube melt down the resistor is supposed to blow before damage occures to the output trans. If there is no chance of tube failure causing damage to the output trans I will remove them as I dont want to add parts that have no purpose.
I would add bias balance adjustment and separate 1...10 ohms cathode resistors to to output tubes to make proper bias adjustment easier.
Actually, I recommend reducing the value of the cathode resistors by a factor of 10. Note this on the outside of the amp, assuming the usual octal connector is used for bias adjustment. Think you will find the bass to magically become more "solid".
I also disable the mono/stereo switch by removing the wires, go with a short twisted pair to the board.
I have never seen a blown output transformer on an ST-70, never even heard on one. I suppose it could happen.
Strapping the tubes for triode is an option that some like, although the output power drops somewhat.
_-_-
I also disable the mono/stereo switch by removing the wires, go with a short twisted pair to the board.
I have never seen a blown output transformer on an ST-70, never even heard on one. I suppose it could happen.
Strapping the tubes for triode is an option that some like, although the output power drops somewhat.
_-_-
Looks like the screen resistors I added are coming out. I have allready bypassed the mono/stereo switch. I made a bucking trans to reduce the input voltage to 114 vac. I also added a 25 ohm thermister on the input side of the power trans. Think Ill try the cathode resistor mod to see if it tightens up the lower end.
Partial cathode coupling on the output stage works well for this amp, see an Audio Research schematic. It does require somewhat more drive.
You can improve the 7199 output's swing symmetry by tuning the 7199 pentode's 1.5M screen resistor for 110VDC at the 7199 triode's pin 8 (cathode).
I used a 1M cermet pot for this, with a 499k series resistor, in place of the original 1.5M resistor.
Last edited:
By reducing the cathode resistors are you saying replacing the 15.6 ohm resistor between the cathode and ground with a 1.56 ohm resistor ?
Check your resistor values on the board, in particular the ones that should be matched. I also add the screen grid resistors based on this article: Techniques To Maximize Power Tube Life, and a quad of KT77s I had that held bias better with the screen stoppers than without.
Yes, around 1 Ohm in EACH tube's cathode to ground is about the smallest you can go. No need for the 1.56 value.
The 15.6 was to make the bias voltage equal to the voltage of a fresh battery for comparison on cheap meters.
Now the bias is 1 Ohm x 50 mA = 50 mV for EACH output tube.
Last edited:
A mod that was championed by AVA (Frank Van Alstine) back in the '80s was to bandwidth limit the input to within the capabilities of the OPT, and increase the value of the interstage coupling caps so that the HPF formed by the coupling caps and grid leak resistors to the EL34s was no longer inside the feedback loop. I think he made the series C at the input 0.022uF with the interstage coupling caps increased to 1uF. He installed a LPF of -3dB at 17kHz at the input too. The article is reprinted here: Our Famous Original Dyna St-70 Rebuild Plans By Frank Van Alstine
I think the low frequency HPF at the input is a good idea, since the amp trying to reproduce 20 to 30 Hz is probably not going to work out very well. I don't know what I think of the LPF at 17kHz. That seems awfully low. But then again, I can only hear to about 13kHz now, so it probably wouldn't matter to me at this point. I noticed it when I was in my twenties and did this mod to an ST70. I think I changed the pole to something like -3dB at 30kHz.
--
I think the low frequency HPF at the input is a good idea, since the amp trying to reproduce 20 to 30 Hz is probably not going to work out very well. I don't know what I think of the LPF at 17kHz. That seems awfully low. But then again, I can only hear to about 13kHz now, so it probably wouldn't matter to me at this point. I noticed it when I was in my twenties and did this mod to an ST70. I think I changed the pole to something like -3dB at 30kHz.
--
Yes, filtering out infrasonic noise out at the amp's I/Ps is a VERY good idea. Doing so protects against O/P trafo core saturation.
IMO, the way to deal with HF error correction signal induced slew limiting is to use high gm small signal tubes. A "simple" way to achieve this end is the installation of a Triode Electronics replacement driver board that's populated with 2X EF86s and an ECC99. The OEM circuit topology is retained, but better small signal active elements are employed. BTW, the 7199 has gone the way of the Dodo and other suitable pentode triodes are becoming extinct too. 😡
While having considerable merit, adding individual bias trim pots. requires "butchering" the chassis. FWIW, a quite decent method of control is available that needs only 1 bias trim pot./channel. You switch to SS rectified B+ and take advantage of the increase in B+ rail voltage by employing combination biased "finals". "Stand" each O/P tube pair on a RC bias network made of a 100 Ω resistor and a 470 μF. capacitor. The 100 Ω resistor provides a convenient "idle" current test point, "eats" the extra B+ volts, and dynamically compensates for minor cathode current differences in the O/P tube pair.
No matter what, the POS C-354 filter choke has got to go. "Shoehorn" a Hammond 159T in.
IMO, the way to deal with HF error correction signal induced slew limiting is to use high gm small signal tubes. A "simple" way to achieve this end is the installation of a Triode Electronics replacement driver board that's populated with 2X EF86s and an ECC99. The OEM circuit topology is retained, but better small signal active elements are employed. BTW, the 7199 has gone the way of the Dodo and other suitable pentode triodes are becoming extinct too. 😡
While having considerable merit, adding individual bias trim pots. requires "butchering" the chassis. FWIW, a quite decent method of control is available that needs only 1 bias trim pot./channel. You switch to SS rectified B+ and take advantage of the increase in B+ rail voltage by employing combination biased "finals". "Stand" each O/P tube pair on a RC bias network made of a 100 Ω resistor and a 470 μF. capacitor. The 100 Ω resistor provides a convenient "idle" current test point, "eats" the extra B+ volts, and dynamically compensates for minor cathode current differences in the O/P tube pair.
No matter what, the POS C-354 filter choke has got to go. "Shoehorn" a Hammond 159T in.
This will overvoltage the main filter capacitor. The rectifier tube is a feature, not a bug.
Or abandon the topology & go with (i recommend) something like Classic Valve Dynamutt. (about half-way down the page)
dave
I too was worried about the high wall plug voltage of 2014, but took the path of putting two 5 W 47 ohm flameproof resistors between the red transformer wires and the plates of the gZ-34 rectifier. These drop the turn on surge current, the running B+ voltage, and provide some protection for the transformer if a rectifier tube shorts out. Lots more 5AR4 tubes short these days than previously, to read the posts on this forum. I put the resistors on cinch solder terminal strips, as I did the grid negative bias silicon diode. Your diodes connected series the 5AR4 may act as a fuse in case of a shorted rectifier tube, I don't know. But a slow enough turn on thermistor allows one the opportunity to think about eliminating the rectifier tube entirely. I'm using 450 VDC radial lead 2000 hour life capacitors under the deck instead of a 525 v rated tall can capacitor.
Don't suppose a 1n4007 will hold off 525 VAC in the long run. They are too short. High voltage can short across dust in a humid environment, particularly as the dust builds up over the years. Salvage a couple of long package (1") high voltage silicon rectifiers with the plastic ring extender from an old CRT terminal or television or somewhere. I don't know where to buy these long arc length diodes in small quantities, they are not popular anymore. Of course I ran my ST-70 >2000 hours per year for a dozen years where dirt (and cat hair) became a significant problem. If this is just an experiment, don't worry about the dust so much.
Oh, BTW, metal film plate resistors does cut the idle hiss a little. I used 2 or 3 w multicomp or vishay. both from an Indian factory. I replaced all the carbon comp resistors 100k or over. The paper caps were replaced with polyester, too, although the original paper ones never failed. The mica capacitor is original, looking like a color domino tile in there.
Last edited:
Sheldon Stokes' cap. board takes good care of that. 😀 Also, the old can caps. are in very bad shape.
BTW, a CL-90 inrush current limiting thermistor between the rectifying SS diodes and the PSU filter provides sufficient delay in B+ rise. The bias supply is "instant" on, which electrostatically shields the O/P tube cathodes.
- Status
- Not open for further replies.