ARC SP-6A Tube Preamp
- Thread starter maddmaxxdrummer
- Start date
I own an SP8 MkII which is not dissimilar.
I have heavily modified/rebuilt it mostly with better capacitors and resistors and changed the MM phono stage for some more gain. I have tried a lot of preamps over time and these old ones rebuilt simply sound better. The gain is IMO part of the reason they have some much energy and life compared to more modern pre-amps. My friend has an SP10 that I used to own and a stock SP10 is easily bettered with a refreshed SP8, I have modified his too with better components and he loves it. For me this is the last preamp I will own.
I have a lot of learning I can share and have worked with Shirokazu Yazaki on these improvements.
Where are you based?
I have heavily modified/rebuilt it mostly with better capacitors and resistors and changed the MM phono stage for some more gain. I have tried a lot of preamps over time and these old ones rebuilt simply sound better. The gain is IMO part of the reason they have some much energy and life compared to more modern pre-amps. My friend has an SP10 that I used to own and a stock SP10 is easily bettered with a refreshed SP8, I have modified his too with better components and he loves it. For me this is the last preamp I will own.
I have a lot of learning I can share and have worked with Shirokazu Yazaki on these improvements.
Where are you based?
What an excellent thread, lots of useful information, and really clear pictures. I'm finding it difficult finding good, clear pictures of old ARC preamps.
I've just bought a half working SP-8, the line stage works but not the phono stage. However, I have managed to diagnose that two of the polystyrene caps in the RIAA feedback stage were short circuit and all the Zener diodes in the PSU's Zener stack were open circuit, which probably over voltaged the RIAA caps. This S/C took out the cathode resistor on one side of the MM stage's first valve. I'm waiting for parts now to complete the build...my first ARC piece of hifi, quite excited.
The preamp looks pretty much untouched, apart from the PSU caps being replaced in the 24V section.
I've just bought a half working SP-8, the line stage works but not the phono stage. However, I have managed to diagnose that two of the polystyrene caps in the RIAA feedback stage were short circuit and all the Zener diodes in the PSU's Zener stack were open circuit, which probably over voltaged the RIAA caps. This S/C took out the cathode resistor on one side of the MM stage's first valve. I'm waiting for parts now to complete the build...my first ARC piece of hifi, quite excited.
The preamp looks pretty much untouched, apart from the PSU caps being replaced in the 24V section.
Attachments
I'd be really interested in to what mods you made to your SP-8 please.
I've always wanted to try an SP-8, so I'd like to squeeze the most juice I can from the one I have.
Back when I was a tech the ARC zener string was infamous for being blown. They ran it hot so the zeners failed, leading to all sorts of other problems. That was by far the biggest issue I recall other than bad tubes.
I was going to mount the replacement zeners slightly off the board, a centimetre or so, to get a little air flow to them.
You might also consider running them at a lower current. Its likely just a single resistor to change.
I like both these ideas. I do not recall the SP-8 specifically, I had moved on by then, but the SP3x and SP6x (as well as their power amps) ran the zeners at pretty high bias even when a pass transistor was used for lower supply impedance.
For a few years I used a series pass regulator identical to the SP-3a1 and the zeners do run hot. When I rebuilt my power supply, I mounted the zeners some distance off the circuit board, maybe 5/8”, and I made sure there was good ventilation in the chassis. I don’t recall any problems with zeners failing. I used 5 watt zeners each time.
You could probably run 1 Watt Zeners with no problems if the bias current for the Zeners and pass transistor were worked out properly.
Thank you for the replies.
I assume that the biasing resistor for the zeners is R66 in the circuit above. There was a thread over of DIY audio where a repair technician increased R66's value from 150k to 220k to take some pressure off the zeners. Interestingly in my SP-8 the CC resistor has naturally increased in resistance to 172k. I'll look to see if I have something a little higher value at 2W.
I'll be using 1W zeners, with a bit of elevation
I assume that the biasing resistor for the zeners is R66 in the circuit above. There was a thread over of DIY audio where a repair technician increased R66's value from 150k to 220k to take some pressure off the zeners. Interestingly in my SP-8 the CC resistor has naturally increased in resistance to 172k. I'll look to see if I have something a little higher value at 2W.
I'll be using 1W zeners, with a bit of elevation
9*47 V = 423 V, but the schematic indicates 394 V, thus lower voltage for the zener string and additional current draw by the circuit (V7 etc.) The nominal drop across R66 is 236 V (current ~1.6 mA), and power dissipated is about 375 mW, well under the 2 W rating. The zeners are rated for 1 W and dissipate <70 mW each (don't know current split between zener string and V7). While that should be acceptable, the carbon resistors ARC used there (IIRC) will indeed rise in value over time due to heat, and there are all those tubes around to increase the temperature even more... That said, these are lower numbers than I remember for the SP3/SP6 but it was a long time ago. I used to replace the bias resistor with 5 W and sometimes used 5 W zeners though am not sure the latter helped much. On my SP3 I added a soft-start circuit to reduce the inrush current into the string but again am not sure if it helped.
R66 also provides plate current for the shunt regulator, V8. Since that's a 12AX7 it does not need a lot of current but both plates are involved so the 150K is like two 300K resistors in parallel. That's right at the upper limit of what works for 12AX7 so I'd leave the 150K value alone. If it were me I'd be running a much lower current through the Zener string to keep it reliable (and not cook the circuit board) but the way this circuit is set up doesn't give you much flexibility. Its a brute force approach where the tube current and the Voltage that is thus present on the plate is also regulated by the Zener string. There's no way you're going to avoid heat as these elements are not exactly working together.Thank you for the replies.
View attachment 142781
I assume that the biasing resistor for the zeners is R66 in the circuit above. There was a thread over of DIY audio where a repair technician increased R66's value from 150k to 220k to take some pressure off the zeners. Interestingly in my SP-8 the CC resistor has naturally increased in resistance to 172k. I'll look to see if I have something a little higher value at 2W.
I'll be using 1W zeners, with a bit of elevation
If it were me I'd probably play with the idea of a Constant Current Source diode in series with R66 and reduce the value of R66 to 75K or so (so the Voltage limit of the CCS diode isn't exceeded). I would then just not use the 12AX7 at all- just unplug it. What is happening here is the 12AX7 and Zener string are regulating the grid Voltage of the pass regulator. But if you use a CCS diode you can get a rock solid Voltage from the Zener string without having to run them so hot. The output Voltage of course would be slightly lower. The CCS diode at the link is good for about 2.42mA. This would reduce the dissipation of the Zener string to about 1/9th to 1/10 of a Watt so they would run nice and cool. The only tricky bit about this idea is the value of R66. CCS diodes can only handle 100V across them so you don't want to come near that limit hence the 75KOhm. I think this would be more reliable and would regulate just as well.
Interestingly the board area where the Zener diodes are doesn't look to have suffered any serious heating. In fact the whole board is pretty free of an heat damage.
Though strange that all bar one Zener was open circuit, and the other being short circuit, wasn't over heating...
Though strange that all bar one Zener was open circuit, and the other being short circuit, wasn't over heating...
It’s strange that nearly all of the Zeners were open circuits. That suggests a single incident causing a major current surge. A power line surge due to a lightning strike or something similar. After all, the Zeners are in series so if one opens up the current through the string goes to zero.
What I saw repeatedly was one zener shorting taking out the entire string with the others usually open.
For completeness this is the link to the DIY audio thread, page 6 that discusses the value of R66.
www.diyaudio.com
ARC SP8 tube reg rebuild
To lighten the stress on the zener string, one could increase the value of R66. I also trimmed R68 to get the U2 output voltage to within the value designated on schematic. This stopped the overheating of V8. I did notice some instability after a while, with U2 output toggling at subsonic...
This makes sense.For completeness this is the link to the DIY audio thread, page 6 that discusses the value of R66.
ARC SP8 tube reg rebuild
To lighten the stress on the zener string, one could increase the value of R66. I also trimmed R68 to get the U2 output voltage to within the value designated on schematic. This stopped the overheating of V8. I did notice some instability after a while, with U2 output toggling at subsonic...
Zeners are usually pretty reliable. I've yet to see one fail on its own if it was operated conservatively.
CCS diodes like the one I linked earlier didn't exist when this preamp was first built. But they do now and have been around for at least 30 years. By using one you kill several birds with one stone; lower temps since there's no 12AX7 tube anymore (good for the power transformer) and the Zener stack will run at a lower temperature, plus its an inexpensive solution. FWIW we use 47V Zeners in diode strings a lot and have done so for the last 40 years.
Well a little bit of progress, and a backwards step..
Coupled with two strange measurements, the 402v B2+ rail into the line stage is not reaching 402V, only to 384V and the voltage after the anode resistors of V4 and V5, the ECC83s in the line stage are oscillating between 174V to 182V at a period of about 10 secs. Cathode bias voltages for V5, V4 and V6 are all correct.
Also, the grid voltage on V6 6DJ8 in the line stage wanders around with the same period as the anode voltage swings.
On the audio output with nothing connected to the input sees the odd voltage blip and spike of 10mVs or so with no real pattern. Sound wise these blips sound like splats and coughs.
It does sound like the PSU on the 402V rail is messing about, certainly not coming up to voltage.
Meanwhile, the voltages to the MM stage are stable before and after the anode resistors.
I'm considering that the diodes and zeners in the PSU are suspect and quite old.
Any pointers would be greatly appreciated.
I've replaced the Zener stack and R66, the 150k resistor at the top of the stack. However, I'm not measuring 394V on the anode of V8, I'm measuring 354V, which indicates to my simple mind that there is too many volts being dropped across R66, hence the current through R66 is on the high side. Whilst the voltage from the PSU caps is at 667V.
Coupled with two strange measurements, the 402v B2+ rail into the line stage is not reaching 402V, only to 384V and the voltage after the anode resistors of V4 and V5, the ECC83s in the line stage are oscillating between 174V to 182V at a period of about 10 secs. Cathode bias voltages for V5, V4 and V6 are all correct.
Also, the grid voltage on V6 6DJ8 in the line stage wanders around with the same period as the anode voltage swings.
On the audio output with nothing connected to the input sees the odd voltage blip and spike of 10mVs or so with no real pattern. Sound wise these blips sound like splats and coughs.
It does sound like the PSU on the 402V rail is messing about, certainly not coming up to voltage.
Meanwhile, the voltages to the MM stage are stable before and after the anode resistors.
I'm considering that the diodes and zeners in the PSU are suspect and quite old.
Any pointers would be greatly appreciated.
Similar threads
