Hi Folks,
Well here's my first post after being registered here for ages!
A good friend recently donated their broken Adcom GFA 585 and I finally have the time available to fix it*. According to the owners' manual it is rated for 250 W into 8 ohms, and 400 into 4 Ohms, with almost monoblock construction - only the power transformer is shared, and that is a real beastie. I also have a GFA 535 that came along for the ride, so a bi-amped setup could be realised.
So, the theme here is big HiFi on a budget (amp has cost me $50 so far). Assuming the restoration is sucessfull, the next step is to find speakers to partner this monster. I'd love to make the most of the current available to drive some Magneplanars, but am happy to consider your recommendations. The listening room is at least 20x20 feet and has sloping ceiling and non parallel walls, no carpet.
Cheers & thanks for a great forum,
Ed
*It looks like an electrolytic capacitor has leaked (seemingly a recurrent problem with this brand) over the input boards - this has disabled one channel which now has a massive DC offset. Hopefully it will be a straightforward clean-up and fix, since no devices appear to be damaged. I have the service manual on order for the sake of completeness.
Well here's my first post after being registered here for ages!
A good friend recently donated their broken Adcom GFA 585 and I finally have the time available to fix it*. According to the owners' manual it is rated for 250 W into 8 ohms, and 400 into 4 Ohms, with almost monoblock construction - only the power transformer is shared, and that is a real beastie. I also have a GFA 535 that came along for the ride, so a bi-amped setup could be realised.
So, the theme here is big HiFi on a budget (amp has cost me $50 so far). Assuming the restoration is sucessfull, the next step is to find speakers to partner this monster. I'd love to make the most of the current available to drive some Magneplanars, but am happy to consider your recommendations. The listening room is at least 20x20 feet and has sloping ceiling and non parallel walls, no carpet.
Cheers & thanks for a great forum,
Ed
*It looks like an electrolytic capacitor has leaked (seemingly a recurrent problem with this brand) over the input boards - this has disabled one channel which now has a massive DC offset. Hopefully it will be a straightforward clean-up and fix, since no devices appear to be damaged. I have the service manual on order for the sake of completeness.
K-amps,
Thanks for the reply. I'd have to check on the suitability of the caps. The PSU is fine though. The problem appears to be 220 uF decoupling caps, mounted on the input boards. These have leaked all over the servo & voltage reference section on the left channel. A little trawling of the net reveals this to be a known problem with the 585 and 565 amp models that use this board. The spilled electrolyte caused complete upset, resulting in the left output being driven fully negative, which blew the O/P fuse. Sadly this was too slow to prevent damage to the speakers belonging to the previous owner.
Anyway, I plan to give both input boards a thorough clean, as there are visual signs of the same problem on the board for the right channel. Then all the electrolytics on these boards will be replaced - just to be safe. Next a trip will to the surplus electronics store for the parts to build a dummy load for this monster before I do a full setup and soak test.
Perhaps one day I'll even get to listen to some music
Cheers,
Ed
Thanks for the reply. I'd have to check on the suitability of the caps. The PSU is fine though. The problem appears to be 220 uF decoupling caps, mounted on the input boards. These have leaked all over the servo & voltage reference section on the left channel. A little trawling of the net reveals this to be a known problem with the 585 and 565 amp models that use this board. The spilled electrolyte caused complete upset, resulting in the left output being driven fully negative, which blew the O/P fuse. Sadly this was too slow to prevent damage to the speakers belonging to the previous owner.
Anyway, I plan to give both input boards a thorough clean, as there are visual signs of the same problem on the board for the right channel. Then all the electrolytics on these boards will be replaced - just to be safe. Next a trip will to the surplus electronics store for the parts to build a dummy load for this monster before I do a full setup and soak test.
Perhaps one day I'll even get to listen to some music
Cheers,
Ed
Well, wish me luck... I'll post my progress here.
An effective cleaning technique is going to be the key thing, as the electrolyte is vile, persistent stuff. In one part of the board, I can get readings anywhere down to around 1 megohm just touching the fiberglass! No wonder the biasing was upset...
I just nipped out during my lunch hour to buy replacement caps, and bits for a dummy load, but am awaiting the service manual in the post before.
Cheers,
Ed
An effective cleaning technique is going to be the key thing, as the electrolyte is vile, persistent stuff. In one part of the board, I can get readings anywhere down to around 1 megohm just touching the fiberglass! No wonder the biasing was upset...
I just nipped out during my lunch hour to buy replacement caps, and bits for a dummy load, but am awaiting the service manual in the post before.
Cheers,
Ed
Progress report.
Over the last couple of evenings, I pulled out both input boards, and removed all the electrolytics. The 220 uF 25 V (3 per board) were the obvious failures, but I removed the two 100 uf 100 V for the sake of completeness. The left channel had large quantities of electrolyte spread across its top surface.
The boards were scrubbed with neat "Simple Green" and rinsed thoroughly in running water, then dried with a hairdryer. After this they had a near new appearence. REplacement caps were soldered in and the boards replaced.
I also made up a dummy load for each channel, with about 80 W handling capacity, ready for testing. All was connected up and power applied*
The result? SUCCESSS!
It was late, so there was only time for brief measurements (no listening). Both channels measured at <20 mV DC offset and very little AC noise, as judged by my Fluke meter. Both channels responded to inputs and all seems well.
K- amps - I would definitely inspect your Adcoms for leaky caps - from what little I have found on the internet and one Adcom approved service dealership I spoke to (They refused to take the amp in for repair!), this appears to be a recurrent problem with the 585 and 565 models - they use these same input boards.
Most of the work time was taken up making careful notes and labeling of the connections to each board - There are some connectors, but also a handful of leads soldered directly with not much spare room to work.
I can't wait for a listening test now
*DO NOT run the unit up on a variac - there's a surge control circuit built in to the AC line circuit. The service manual warns that a resistor in this circuit will burn out if power is applied gradually.
Over the last couple of evenings, I pulled out both input boards, and removed all the electrolytics. The 220 uF 25 V (3 per board) were the obvious failures, but I removed the two 100 uf 100 V for the sake of completeness. The left channel had large quantities of electrolyte spread across its top surface.
The boards were scrubbed with neat "Simple Green" and rinsed thoroughly in running water, then dried with a hairdryer. After this they had a near new appearence. REplacement caps were soldered in and the boards replaced.
I also made up a dummy load for each channel, with about 80 W handling capacity, ready for testing. All was connected up and power applied*
The result? SUCCESSS!
It was late, so there was only time for brief measurements (no listening). Both channels measured at <20 mV DC offset and very little AC noise, as judged by my Fluke meter. Both channels responded to inputs and all seems well.
K- amps - I would definitely inspect your Adcoms for leaky caps - from what little I have found on the internet and one Adcom approved service dealership I spoke to (They refused to take the amp in for repair!), this appears to be a recurrent problem with the 585 and 565 models - they use these same input boards.
Most of the work time was taken up making careful notes and labeling of the connections to each board - There are some connectors, but also a handful of leads soldered directly with not much spare room to work.
I can't wait for a listening test now
*DO NOT run the unit up on a variac - there's a surge control circuit built in to the AC line circuit. The service manual warns that a resistor in this circuit will burn out if power is applied gradually.
The service manual makes for some interesting reading. These are some of the points that made me sit up and take notice, and may be of interest to others:
Whilst the topology is essentially a differential input stage (allowing for feedback application), voltage gain and emitter follower complementary output stage, there is some real sophistication in the design.
Quote "The amplifier employs a descrete dual-differential cascode class A input stage followed by a dual cascode class A voltage gain stage..."
Quote "The output stage is capable of greater than 60 amps into low impedance loads."
With all this current available, the biasing networks of these stages cleverly incorporates thermal protection and power supply failure protection to shut the amp down, protecting the output and, presumably anything connected to it. This did not save my friends speakers when the circuit was swamped in electrolytic capacitor juice!
A DC servo OPAMP is used to correct imbalance in the output stage.
The input stage has a freq response 0.7 Hz to 700 kHz, with roll off applied for an actual -3 dB point at 80 kHz.
Some selected (factory) specs:
Power output 250 W into 8 Ohms, <0.02% THD
Power output 400 W into 4 Ohms, <0.02% THD
@1W into 8 Ohms, frequency response of 10 Hz to 20 kHz =), -0.5 dB
20% to 80% rise time for 120V p-p 5kHz square wave, 2.9 microseconds
Damping factor (20 Hz to 20 kHz) >=600
Weight, 45 lbs
...I know, I mistakenly took a flight with this thing as checked baggage.... kerching! Overweight bag...
Of course, none of this tells me what it will sound like - I'm still in the dark here. At least I'm in a position to go speaker shopping now...
Cheers,
Ed
Whilst the topology is essentially a differential input stage (allowing for feedback application), voltage gain and emitter follower complementary output stage, there is some real sophistication in the design.
Quote "The amplifier employs a descrete dual-differential cascode class A input stage followed by a dual cascode class A voltage gain stage..."
Quote "The output stage is capable of greater than 60 amps into low impedance loads."
With all this current available, the biasing networks of these stages cleverly incorporates thermal protection and power supply failure protection to shut the amp down, protecting the output and, presumably anything connected to it. This did not save my friends speakers when the circuit was swamped in electrolytic capacitor juice!
A DC servo OPAMP is used to correct imbalance in the output stage.
The input stage has a freq response 0.7 Hz to 700 kHz, with roll off applied for an actual -3 dB point at 80 kHz.
Some selected (factory) specs:
Power output 250 W into 8 Ohms, <0.02% THD
Power output 400 W into 4 Ohms, <0.02% THD
@1W into 8 Ohms, frequency response of 10 Hz to 20 kHz =), -0.5 dB
20% to 80% rise time for 120V p-p 5kHz square wave, 2.9 microseconds
Damping factor (20 Hz to 20 kHz) >=600
Weight, 45 lbs
...I know, I mistakenly took a flight with this thing as checked baggage.... kerching! Overweight bag...Of course, none of this tells me what it will sound like - I'm still in the dark here. At least I'm in a position to go speaker shopping now...
Cheers,
Ed
I have been going nuts trying to get the DC offset down on a 585. I have cleaned the boards with a conservative amount of alcohol & flux remover, but I'm going to try this simple green and a more vigorous cleaning.
Had the same problem as Ed Holland - leaky bypass caps (also change the VBE multiplier cap as it's the same type/brand), blown zener & DC servo. All replaced, but I still have an offset of .7v and 1.2v.
Sound-wise, the 585 has the Adcom 5xx characteristic but more refined over the 555 or 555II across all frequencies with excellent speaker control and sweeter top end. The 565 monos have more guts, but also have twice the output devices. This amp is an excellent chioce for Maggies.
Cheers,
-JamesW-
Had the same problem as Ed Holland - leaky bypass caps (also change the VBE multiplier cap as it's the same type/brand), blown zener & DC servo. All replaced, but I still have an offset of .7v and 1.2v.
Sound-wise, the 585 has the Adcom 5xx characteristic but more refined over the 555 or 555II across all frequencies with excellent speaker control and sweeter top end. The 565 monos have more guts, but also have twice the output devices. This amp is an excellent chioce for Maggies.
Cheers,
-JamesW-
Woodman:
I'm certain that this problem calls for really thorough cleaning. The electrolyte is water soluble, but the detergent helps too. Scrubbing a poplulated PCB under running water seems wrong, really wrong, but is actually fine if it is dried quickly and thoroughly afterward. I did remove the bias trim pot beforehand.
Regarding bias, the service sheet calls for a voltage measurement between test points (TP301 & 401), one on each of the two transistor boards for each channel. Adjustment is made (no input) for a 24 +/-1 mV reading after 10 minute warm up. Then a running period of 80W output for 10 minites, followed by readjustment.
Strangely, there is no mention of offset voltage specification, or acceptable limits.
Last night I hooked the amp up to some speakers, after running into a dummy load for a while to prove that it can behave itself. Though the speakers are certainly not the last word (Sony 3 way design from the 80s) the output was clean and quiet.
Cheers,
Ed
I'm certain that this problem calls for really thorough cleaning. The electrolyte is water soluble, but the detergent helps too. Scrubbing a poplulated PCB under running water seems wrong, really wrong, but is actually fine if it is dried quickly and thoroughly afterward. I did remove the bias trim pot beforehand.
Regarding bias, the service sheet calls for a voltage measurement between test points (TP301 & 401), one on each of the two transistor boards for each channel. Adjustment is made (no input) for a 24 +/-1 mV reading after 10 minute warm up. Then a running period of 80W output for 10 minites, followed by readjustment.
Strangely, there is no mention of offset voltage specification, or acceptable limits.
Last night I hooked the amp up to some speakers, after running into a dummy load for a while to prove that it can behave itself. Though the speakers are certainly not the last word (Sony 3 way design from the 80s) the output was clean and quiet.
Cheers,
Ed
Hi Ed,
"Strangely, there is no mention of offset voltage specification, or acceptable limits."
There's no mention because it's supposed to be in the uV range due to the matching of the input devices and the DC servo. The real acceptable limit is no more than 30mv.
IMO, the design of this amp is overly complicated - twice the chance for something to go wrong and a huge pain to repair (quickly anyway). The dual differential was to keep noise & distortion low, but the darlington input pairs, the shunt regulator for input bias and servo power was just a bit much I think. No one at Adcom knew how to fix these amps beyond a blown output. DC offset problems were solved by replacing the servos until an acceptable limit was found, otherwise the input board was simply replaced.
The 565 was the brainchild of Victor Campos and executed by Walt Morrey with Walt Jung contributing on the DC servo (that's what Victor said, but this is some 20yrs after the fact). The 585 was trickled down from the 565, and the 555II was trickled down from the 585.
Hi K-Amps,
I think we have different sonic tastes. Yes, the mids are a bit analytical, but I don't regard analytical as a bad thing. I find the mids so much more clean than the other amps (555/555II, etc) and with a better presentation than any of the MOSFET Adcoms.
Regards,
-JamesW-
"Strangely, there is no mention of offset voltage specification, or acceptable limits."
There's no mention because it's supposed to be in the uV range due to the matching of the input devices and the DC servo. The real acceptable limit is no more than 30mv.
IMO, the design of this amp is overly complicated - twice the chance for something to go wrong and a huge pain to repair (quickly anyway). The dual differential was to keep noise & distortion low, but the darlington input pairs, the shunt regulator for input bias and servo power was just a bit much I think. No one at Adcom knew how to fix these amps beyond a blown output. DC offset problems were solved by replacing the servos until an acceptable limit was found, otherwise the input board was simply replaced.
The 565 was the brainchild of Victor Campos and executed by Walt Morrey with Walt Jung contributing on the DC servo (that's what Victor said, but this is some 20yrs after the fact). The 585 was trickled down from the 565, and the 555II was trickled down from the 585.
Hi K-Amps,
I think we have different sonic tastes. Yes, the mids are a bit analytical, but I don't regard analytical as a bad thing. I find the mids so much more clean than the other amps (555/555II, etc) and with a better presentation than any of the MOSFET Adcoms.
Regards,
-JamesW-
Woodman,
Thank you very much for the background information - most interesting.
I agree that the design is complicated - reading the schematic is quite a challenge!
When all is warmed up, my amp has about 2 mV offset on the right channel, and perhaps 10 on the left. I have spent quite a little time monitoring this, especially considering the gross failure in the left channel.
From switch-on, there is about 200 mV offset in each channel. This falls quickly as the circuit settles. In the right channel, offset gradually approaches the steady state value (2 mV). In the left channel, the offset swings through, zero reaches about - 60 mV, then returns gradually to 10 mV.
I'm unsure what to read into this, or whether I should pay any attention to it at all - any thoughts?
The other thing that bugs me somewhat is the offset control IC. It's clearly a single op-amp device with a standard pin-out, but is marked with an Adcom part number (Adcom 2A). I searched and found another thread HERE which suggested that the AD820 is appropriate, and the requirements are a low power device that will behave properly if input voltages reach supply rail values.
Any thoughts on this anyone?
Thanks again for everyone's input to this thread - it has rather deviated from the original question regarding speakers though
Cheers,
Ed
Thank you very much for the background information - most interesting.
I agree that the design is complicated - reading the schematic is quite a challenge!
When all is warmed up, my amp has about 2 mV offset on the right channel, and perhaps 10 on the left. I have spent quite a little time monitoring this, especially considering the gross failure in the left channel.
From switch-on, there is about 200 mV offset in each channel. This falls quickly as the circuit settles. In the right channel, offset gradually approaches the steady state value (2 mV). In the left channel, the offset swings through, zero reaches about - 60 mV, then returns gradually to 10 mV.
I'm unsure what to read into this, or whether I should pay any attention to it at all - any thoughts?
The other thing that bugs me somewhat is the offset control IC. It's clearly a single op-amp device with a standard pin-out, but is marked with an Adcom part number (Adcom 2A). I searched and found another thread HERE which suggested that the AD820 is appropriate, and the requirements are a low power device that will behave properly if input voltages reach supply rail values.
Any thoughts on this anyone?
Thanks again for everyone's input to this thread - it has rather deviated from the original question regarding speakers though
Cheers,
Ed
Hi Ed,
The 820 is not the op-amp to use! The 820 is ok for the 555II because it uses a monopolar supply for the servo. The 585/565 have a bipolar supply so it's better to use something like the AD711 like k-Amps suggests. I forgot what the Adcom 2A really is, I had it written down somewhere.
The changing offset is normal just because you have three circuits affecting offset coming up to their quiescent points. Generally Adcom amps take around 45min to settle in.
For your room you can drive just about any loudspeaker you want. Magnepans, Thiels, Von Schweikerts, Vandersteens, B&W, KEF, etc. any large full range will work. The amp has a clean, slightly analytical sound, but you can get good sound/synergy in your room by assuming the amp doesn't exist and matching the characteristic of your preamp with that of the speakers.
Cheers,
-JamesW-
The 820 is not the op-amp to use! The 820 is ok for the 555II because it uses a monopolar supply for the servo. The 585/565 have a bipolar supply so it's better to use something like the AD711 like k-Amps suggests. I forgot what the Adcom 2A really is, I had it written down somewhere.
The changing offset is normal just because you have three circuits affecting offset coming up to their quiescent points. Generally Adcom amps take around 45min to settle in.
For your room you can drive just about any loudspeaker you want. Magnepans, Thiels, Von Schweikerts, Vandersteens, B&W, KEF, etc. any large full range will work. The amp has a clean, slightly analytical sound, but you can get good sound/synergy in your room by assuming the amp doesn't exist and matching the characteristic of your preamp with that of the speakers.
Cheers,
-JamesW-
Hey K-Amps,
6v sounds a bit extreme, but do the wash first and see what happens.
Remove all the electrolytics and the bias pots. Run under warm water and then spray Simple Green. Work it in carefully with a paintbrush on the component side and then rinse. Shake off excess, pat with paper towel. I used compressed air at this point to remove as much moisture as possible before drying with the heat gun set on low. Get the board nice and warm, but not so hot to melt insulation. Make double sure it's dry. I also cleaned up the bottoms of the board with isopropanol.
After reassembly and power up I got -3mv DC on both channels. I could practically kiss Ed for this as I was totally baffled after replacing/swapping/measuring every component that could contribute to offset. I repaired the amp the first time around, but that pesky electrolyte was still on the board even after conventional cleaning.
If the wash does not bring your offset down to an acceptable level, then it's either the servo, the input bias/servo supply or the input pairs have drifted. These are a PITA to match because you have to build a circuit to match them and then sit like a moron watching the meter settle. PM me if you need input pairs.
Ed,
Dude, you RAWK!
Cheers,
-JamesW-
6v sounds a bit extreme, but do the wash first and see what happens.
Remove all the electrolytics and the bias pots. Run under warm water and then spray Simple Green. Work it in carefully with a paintbrush on the component side and then rinse. Shake off excess, pat with paper towel. I used compressed air at this point to remove as much moisture as possible before drying with the heat gun set on low. Get the board nice and warm, but not so hot to melt insulation. Make double sure it's dry. I also cleaned up the bottoms of the board with isopropanol.
After reassembly and power up I got -3mv DC on both channels. I could practically kiss Ed for this as I was totally baffled after replacing/swapping/measuring every component that could contribute to offset. I repaired the amp the first time around, but that pesky electrolyte was still on the board even after conventional cleaning.
If the wash does not bring your offset down to an acceptable level, then it's either the servo, the input bias/servo supply or the input pairs have drifted. These are a PITA to match because you have to build a circuit to match them and then sit like a moron watching the meter settle. PM me if you need input pairs.
Ed,
Dude, you RAWK!
Cheers,
-JamesW-
Woodman - again, that's really excellent info, thanks. And also the recommendation on speaker - preamp matching.I realise now that it's almost a stupid question to ask what speakers to match with an amp like this - it will drive almost anything! The thing is I have never owned anything in this league before, so was looking for pointers. My other amps are 40 and 50 W valve (tube) and transistor amps respectively, and fluffy kittens compared to this Jaguar.
In fact, I rather fancy the challenge of designing a straightforward preamp. There's lots of opportunity to play around with circuits.
I understand this, but actually took another look at the 585's left channel board, and scrubbed the track side again over the weekend, removing the laquer, but not the green solder mask. It seems there was some contamination infused within it - cap leakage had in fact caused some minor track corrosion . This seems to have cleared things up totally - there's no "wavering" in the offset now, just a steady correction from switch-on, almost identical to the right channel. I'm now happy that the amp is going to behave itself.
Cheers,
Ed
In fact, I rather fancy the challenge of designing a straightforward preamp. There's lots of opportunity to play around with circuits.
I understand this, but actually took another look at the 585's left channel board, and scrubbed the track side again over the weekend, removing the laquer, but not the green solder mask. It seems there was some contamination infused within it - cap leakage had in fact caused some minor track corrosion . This seems to have cleared things up totally - there's no "wavering" in the offset now, just a steady correction from switch-on, almost identical to the right channel. I'm now happy that the amp is going to behave itself.
Cheers,
Ed
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.