SUMO Andromeda II bias or idle power?

[SQLGuy]

Hi folks,

I have a SUMO Andromeda II amp that I've owned for many years, and blew up and repaired about a year ago. When I repaired it, I measured the working channel's source resistor drops and set the bias on the repaired side to match this.

However, the amp runs rather cold, which surprises me for a MOSFET design, and I don't know that the working channel was biased correctly at the time I did the repair.

Anyone know what the bias current or bias setup procedure is for this amp? Also, if you do, would you also happen to know the setup procedure for the NULL adjustment?

Thanks,
Paul

 

[llwhtt]

Track down Andrew Hefley if you can, I think he is responsible for that amp. He was head of Sumo after Sir James. Last I heard he was designing for Monarchy Audio. Other than that I don't know where to go. I myself have a II cadaver, one complete channel was missing when it was given to me. It's being made into a Pass A40 now.

Craig

 

[h_a]

Hi,

here you find a schematic of the Sumo:

http://www.audio-circuit.dk/images/schematics/Sumo-Andromeda-ll-pwr-sch.pdf

While it does not contain hints on biasing it can help you to figure out how to adjust the offset.

Have fun, Hannes

EDIT: I'm in weekend mode, so I didn't immediately thought that this is a servo design, so there shouldn't be any offset adjustments necessary. Do you have offset?

 

[SQLGuy]

I have that schematic, and for me and another friend who designs amps, at least, it did not give any clues as to how to adjust the NULL.

I did read an old post somewhere that suggested that the NULL adjustment has to do with the output not changing in amplitude when the load is changed. Since this is a differential output (bridged) amp, and there are NULL adjustments for each phase (two per channel), I am still not clear on what the designer would have intended as the process for adjusting this, assuming this IS the correct definition of NULL - i.e. how much load, where to ground it so you can check one phase at a time, what kind of a test signal....

 

[llwhtt]

Just got this info from an ex-SUMO tech, "bias is 250ma per channel quiescent and set the distortion null with an audio analyzer for minimum residual notch at crossover". I repaired a II a few weeks ago, before this info and I ended up in the low 300ma area for bias which was minimum distortion at 1 Watt into 8 Ohms then backed off a tad and then I set the null for minimum distortion at 50 Watts into 8 Ohms. Both adjustments were done at 1KHz. Current was measured across the + rail fuse holder, fuse removed of course. While waiting on parts for the dead channel I screwed around with the good channel for hours trying to figure out the null adjustment. Using a Sound Technology analyzer (balanced input) I found that after setting the idle current I could see the distortion was greater at 50 Watts and when I adjusted the null I could reduce the distortion to at least match the 1 Watt setting. Went back and forth a few times and all was good at both levels. I just pulled the 50 Watt level out of my you know what so other levels may work just as well. The other thing I did was to monitor heatsink temperature with an IR thermometer making sure both heatsinks of each channel were the same or close. For those of you that don't know this amplifier there are two bias and null adjustments per channel and no easy way to measure the idle current for each when the amp is assembled. When I first tried to run the good channel it oscillated depending on level and frequency and just looked terrible on the scope. Hope this helps as many of you know this info is beyond scarce.

Craig

 

[SQLGuy]

Thanks Craig, this is useful information.

For the NULL adjustment, did you measure and adjust each phase separately referenced to chassis, or did you adjust the NULLs on both phases to see which made the best result for 50W distortion at the balanced output?

Also, can you tell me what outputs that particular II had? I've seen them with TO-3 Motorolas (MTM20N10 and MTM20P10) and with TO-3P IRs (IRF240 and IRF9240 and others). I've also heard of them using some TO-3 IR's and I would think that more modern IRF9140's and 140's would be reasonable choices as well....

Thanks,
Paul

 

[llwhtt]

Hi Paul,

I used the balanced output into the Sound Technology analyzer which has balanced inputs. After setting the bias and increasing the power to 50 Watts I started tweaking the "nulls" a little at a time going from one to the other until the distortion was the lowest on the meter. You might be able to do one phase at a time using ground as the reference, I didn't try that. Unlike Sir James' designs this is NOT a true balanced circuit it is two amps per channel run in bridged mode so using ground and adjusting one phase at a time may work OK.

This amp came in with the + rail fuse blown on the left channel. The + MOSFETs, drivers, and bias transistors were shot along with the + null LEDs and a couple of resistors. I used matched IRFP240s and 9240 from TECH-DIY. They cost a little more but I didn't have to buy as many and have to match them myself. After the repair and alignment all of the adjustment pots were very close in there relative positions between channels. When I laid one trace over the other on the scope the traces were identical. Even though this amp needed some parts replaced I think the alignment made the most difference as the good channel looked like crap on the scope before I tweaked on it. I don't know if the alignment was bad or the pots were dirty. The pots are not sealed and after twenty years they may have been oxidized.

Another thing I noticed is the amp draws much more current a start up and after being on for a few minutes the idle current settles down. After staring at the PCB for a while I noticed that the bias and driver transistors are very close to each other and the driver runs very warm so I'm thinking as the driver warms up it warms up the bias transistor and therefore the idle current comes down.

As far as the different versions of this amp I know of three, TO3s and NO null LEDs, TO3s and null LEDs, and TO247s and null LEDs. The only schematic floating around is the TO3 and LED version.

Craig