I just acquired a pair of Threshold SA-1s brand new sealed in their original boxes as part of a larger purchase of NOS Threshold gear. Given these are 34 years old or so, is there any harm in bringing them up slowly on a variac one at a time to test? How slowly should I go? Power consumption will likely be around 400W at line voltage. I will likely replace the main filter capacitors any way, but I'd like to confirm the amplifiers work before I get inside of them and start replacing filter capacitors. Any recommendations on filter caps I should use? On paper, EPCOS/TDK seem to be the best ones currently available.
Conventional wisdom sez variac, but I have not seen a case where you couldn't just plug these in. The cans used were pretty rugged.
Thank you for your response, Mr. Pass. If you were to build these today, would you choose the Mallory/CDE caps or another brand?
It took three variacs before I found one that didn't start humming obnoxiously at higher amperage, but the amplifier sounds beautiful after bringing it up slowly with output to an ~8Ω test speaker. The power consumption was initially ~500W but slowly began to fall as the unit warmed up. Settled in below 400W after about 20-30 minutes with the heatsinks looking like this:
Quite warm, but not too hot to leave my hands on the heatsinks for an extended period though I'm sure they will heat up a bit more if left on for longer. With that, as the market is falling to new YTD lows, it's time to get back to work.
Quite warm, but not too hot to leave my hands on the heatsinks for an extended period though I'm sure they will heat up a bit more if left on for longer. With that, as the market is falling to new YTD lows, it's time to get back to work.
Used the variac on the other SA-1 and this was the result:
Appears something has gone out of spec after 34 years. Another interesting thing I've noticed is that the power consumption is significantly lower through the variac when brought up slowly to line voltage (I measured what level on the variac was truly line voltage) than if I plug in the amplifiers directly to the wall and turn them on. On this second amplifier, power consumption fell to around 300 watts before I turned it off likely to compensate for that hot spot on the right. Any ideas? corrections?
Appears something has gone out of spec after 34 years. Another interesting thing I've noticed is that the power consumption is significantly lower through the variac when brought up slowly to line voltage (I measured what level on the variac was truly line voltage) than if I plug in the amplifiers directly to the wall and turn them on. On this second amplifier, power consumption fell to around 300 watts before I turned it off likely to compensate for that hot spot on the right. Any ideas? corrections?
If Jon Soderberg at Vintage Amp Repair in CA is still in business, you might want to get these checked by him. He was a service Tech at Threshold when these amps were made 40 odd years ago. He is the expert.
Yes, I’ve heard of him. He’s about three hours up the valley from me. I wanted to go through this in a deliberate manner so as to maybe settle some long-standing debates around electrolytic capacitors and how they age at least in my own mind. The issue of power consumption seems to me to be because of the filter capacitors.
At 34 years old, these capacitors are likely chemically compromised and no level of electrical therapy will bring them all the way back even if they have very few hours on them. I have an S/300 that had been used quite a lot and the power consumption behaved similarly (started high and fell quite a lot over time). When I replaced the filter capacitors, the issue was resolved. Power consumption did not fluctuate nearly as much with new EPCOS/TDK caps and was overall lower.
Yes, it is nice he is close. Hopefully it does not come to lugging these up there though. Just to recap this pair, it's going to be about $700. Filter caps this size have gotten quite expensive.
If you are going to spend that money, then buy quality caps - a quick phone call to Jon on the subject might be worthwhile if he recommends a certain replacement cap.
For the big cans - check out the prices at Online Components there in your country.
For the big cans - check out the prices at Online Components there in your country.
Rather, I note a severe 'cool spot' front-right with possibly another rear-right.
I bet these have multiple big transistors and one or two (or their wires!) have quit working. Depending on the bias, the surviving parts mat be taking-up the slack. As even the 'hot spot' temps are not too distressing, and Pass didn't build junk, I doubt they died of the heat. Maybe random factory defect, maybe somebody has been messing around inside.
Yes, I think this goes without saying. Thankfully, I am capable of reading data sheets and the options are few at this size of capacitor so I'm going to go with the TDK/EPCOS. This is not a difficult comparison:
Unit #2 (the one with the uneven heatsink temps) has to come apart for a few reasons.
1) For lack of a better word, the unit is "tweaked". It has a small ding on one of the corners that I doubt could've happened in transit given the thick packing material that showed no signs damage. That I could ignore, these units have been floating around in a basement for 34 years after all. However, when I pulled the top cover, I noticed there was a bit of damage around the screw holes, where the screw threads had impacted the edges of the holes and the heatsinks are shifted slightly relative to what I suppose would be their original locations. At least nothing appears to be bent, though I suppose, as a result of this force, there could've been physical damage to the transistors or their adhesion to the heatsinks.
2) The 1Ω emitter resistors appear to have drifted. Some measure 1.1Ω and others 1.2Ω so I'm going to replace them all.
3) I want to make some updates to the driver board(s).
Any advice on how best to get this amplifier apart with the least amount of soldering iron usage?
1) For lack of a better word, the unit is "tweaked". It has a small ding on one of the corners that I doubt could've happened in transit given the thick packing material that showed no signs damage. That I could ignore, these units have been floating around in a basement for 34 years after all. However, when I pulled the top cover, I noticed there was a bit of damage around the screw holes, where the screw threads had impacted the edges of the holes and the heatsinks are shifted slightly relative to what I suppose would be their original locations. At least nothing appears to be bent, though I suppose, as a result of this force, there could've been physical damage to the transistors or their adhesion to the heatsinks.
2) The 1Ω emitter resistors appear to have drifted. Some measure 1.1Ω and others 1.2Ω so I'm going to replace them all.
3) I want to make some updates to the driver board(s).
Any advice on how best to get this amplifier apart with the least amount of soldering iron usage?
PRR told you what's most likely wrong with one of the amps. Based on my experience with thousands of amps I built when I worked for PA amplifier manufacturer, he is most likely right!
You should connect a signal generator and oscilloscope to a good working amp and check when it starts to clip UNLOADED.
Then repeat with the suspect amplifier also UNLOADED. That will tell you if all transistors are working; are biased properly and share the load (quiescent current) equally.
I'd assume the amplifiers are built to provide full output swing even if unloaded without being damaged (they should). Nelson can confirm this....
The difference in measurement of 1-ohm resistors can easily be contributed to variations in contact resistance between probes, how probes are seated in the multimeter sockets and how you attach the probes to that resistor.
The capacitors look good. I do not think you should be replacing them. Nelson could also tell you what ripple he'd expect to see on V+ and V-. If you measure close to that value then you'll know for sure they are okay.
Based on what you wrote so far... maybe it would be prudent to take the amps to that Jon Soderberg fella.
You should connect a signal generator and oscilloscope to a good working amp and check when it starts to clip UNLOADED.
Then repeat with the suspect amplifier also UNLOADED. That will tell you if all transistors are working; are biased properly and share the load (quiescent current) equally.
I'd assume the amplifiers are built to provide full output swing even if unloaded without being damaged (they should). Nelson can confirm this....
The difference in measurement of 1-ohm resistors can easily be contributed to variations in contact resistance between probes, how probes are seated in the multimeter sockets and how you attach the probes to that resistor.
The capacitors look good. I do not think you should be replacing them. Nelson could also tell you what ripple he'd expect to see on V+ and V-. If you measure close to that value then you'll know for sure they are okay.
Based on what you wrote so far... maybe it would be prudent to take the amps to that Jon Soderberg fella.
To suggest immediately that multiple transistors have failed in an amplifier that's never been used is absurd, especially when one considers that no one output transistor is used at more than 5% of its rating. Additionally, I very much doubt the amplifier would behave in this manner if that were the case. What I likely have is an issue of beta matching and perhaps some unevenness due to the emitter resistors.
This is unnecessary. If the current was shared equally, the heatsink wouldn't have such uneven temperature like it does.
Or the resistors are just 34 years old and have drifted.
Please find me documentation where capacitors that have sat unused for 30+ years should be used. You won't. Despite bringing the amplifiers up on the variac first, they're consuming well over 500 watts at power on for several moments without the variac and above 400 watts for 10-20 minutes afterward before they finally settle down just below the 400 watt specification. If I put new capacitors in, I assure you the power consumption would look very different as it did on my S/300 II.
I don't think it would be at this point. These are not very complicated amplifiers and I've dragged multiple members here through successful projects before despite similar protests.
Last edited:
Good luck with the fix however you proceed with it - one thing is for sure, these beauties are worth every cent in restoring to 100% condition.
Have you had a go at measuring the bias (volts across emitter resistors) and the DC offset voltage. Measuring the bias volts across each emitter resistor will lead you to any output transistor problems. Also, you mentioned the difference in emitter resistance values - not sure, but maybe way back then the resistors might have been 10% tolerance, replacing with more modern close tolerance and better resistors will help. More than likely the caps will be OK, I bet NP would say they will be fine.
Have you had a go at measuring the bias (volts across emitter resistors) and the DC offset voltage. Measuring the bias volts across each emitter resistor will lead you to any output transistor problems. Also, you mentioned the difference in emitter resistance values - not sure, but maybe way back then the resistors might have been 10% tolerance, replacing with more modern close tolerance and better resistors will help. More than likely the caps will be OK, I bet NP would say they will be fine.
Last edited:
If it ain't broke don't fix it!!
Check it thorough. Currents voltages offsets PSU ripples under the load. Caps may re-form to near original values when put under tension if not dried.
In regard to resistors: unless you have four wire ohm meter for low values the measurements of yours are, let's say, nor precise, not mentioning oxidation of terminals etc. Assuming welded construction no way a piece of wire will change 10 % in 30 years. The latter is a life span of an airplane; it does not become 10% longer or 10% thinner or does it?
Check it thorough. Currents voltages offsets PSU ripples under the load. Caps may re-form to near original values when put under tension if not dried.
In regard to resistors: unless you have four wire ohm meter for low values the measurements of yours are, let's say, nor precise, not mentioning oxidation of terminals etc. Assuming welded construction no way a piece of wire will change 10 % in 30 years. The latter is a life span of an airplane; it does not become 10% longer or 10% thinner or does it?