Tweakers' Asylum

Posted by DeanKS (A) on January 13, 2006 at 00:01:14

I posted earlier seeking help with mechanicals of this amp. It is very heavy and to work on the output boards you have to remove them from the heat sinks or remove the heat sinks c/w boards. Both techniques are difficult. It does not help that the two output boards have some hard solder connection(s) to each other :(

Summary: PS board- added 8x100uf Blackgate & 8x0.47uF film caps
Output boards- replaced total of 4 100uF with 200uF Blackgate

The 8 output drivers per channel are attached to the heat sinks with hex cap screws and those go through thick strip off aluminum to offset them from the heat sink which provides clearance behind the CB. The board itself is attached to the heat sinks with 6 1/4" nuts that thread onto setscrew style studs that go directly into the heat sink, with short aluminum spacer bushings. There is one other device attached to the separate strip of aluminum that I left attached, so the strip of aluminum stayed with the CB and kept the insulators, strip and drivers all together. But that is a very fragile arrangement and takes much care. Might be better to simply remove its single small Philips screw. The strip also has 6 heavy Philips screws to the heat sink. Do not mix up the heat sinks L-R.

When replacing the covers, they can easily not fit the tops of the heat sinks. What you need to do is loosen the bottom 4 screws per heatsink (each of which is really in two parts), and then get the top four countersunk screws in place, along with all of the smaller ones. Once that is done, then secure the bottom screws which have some degree of freedom that the top ones do not.

Power supply board:
This is easy to remove to work on. I used a digital camera to document wires (about 20) in case I needed help later. But for the most part, the well laced up wire bundles keep all of the wires properly located so that refit is simple.
I found that there was enough room between the PS CB and the top cover to mount 100uF 100VDC Blackgates right onto the top of the CB. I laid out holes for the leads that went through the proper traces on the bottom of the CB. Then scraped off solder resist around each to provide solder pads. Located so that none would be under the footprint of the main PS caps below the CB. Also cleared copper off the the top of the CB where I drilled through parts of the Aragon in-copper logo so that these would not be unexpectedly hot if someone were exploring a hot circuit and did not expect such to be live. Eight 100Uf caps were inserted from the top and soldered in place. Some tidy amounts of hot melt glue to make them mechanically secure.
The CBs are thick and the PS CB drilled very easily. I am thinking that these might not be fiberglass. Perhaps polyester fiber?
Tacked 8 .47uF 250V Philips metalized polypropylene "yellow cube" caps (near the end of my cache from 1980s') on the bottom of the CB. Four across the outputs downstream of the fuses, and 4 across the CB's output fuses and secured with hot melt. The caps across the fuses create a low resistance path for AC/transients on the DC current. If a fuse blows, the cap blocks DC and the amp shuts down as one would expect. (So there is no created risk.)

Transformer:
Pulled the transformer, harness, protection CB and grounding CB to get room to work on the left channel output module. So the amp was very much road kill in appearance for a while. The power connection module is removed with the harness intact... just remove the two small Philips screws and nuts. Four spade connectors need to be removed from the panel switch.

Output boards:
Each had 2 100uF 100V electrolytic caps, one for each rail to ground. That is not much local energy storage for a heavy current amp. These were my main motivation. I removed these. The holes are through plated, so one needs to dry these out with a solder sucker or wick. The traces on the back are thick and wide. A small electronics soldering station is almost useless. I used the big old Weller soldering gun. I replaced those tiny 100uF caps with 200uF 100V Blackgates. These were very large compared to the old caps. But there was still 1/2" clearance to the space limited right channel side cover. The diameter fit into the space available, but only just. I did not want to drill the CB, but could have. The bottom lead went straight in and the upper one got about a 1.5mm tight jog to fit that hole. On the reverse, I turned the soldering pad into a soldering strip, and bent the leads down and soldered a length to get better contact with the buss bar sized power and ground traces. So all that was done on the output boards was the total of 4 200uF caps. From what I had read a long time ago, it was stated that Blackgate caps do not need to be bypassed. There are 2 .22uF film caps on the rails, but that really does not amount to much. (On my Adcom 555 I have 8 .47Uf PP caps per channel for local storage)

I put all back together, cleaned screw terminals and lugs, fuse clips, fuse end caps, main PS cap terminals, contact lands on CD for the main caps and applied Quicksilver Gold to all. Also pulled the voltage selector out and treated its contacts and fuse.

Did the smoke test:
Nothing happened (no smoke), nothing at all, no power in evidence. Then I put the fuse into the voltage selector and it started up with the familiar grunt, pause, speaker out relays pull in. Checked the output terminals (nothing attached) and found that the output offsets were less that 1 mV (undetectable). This step is important after a significant mod, as you need to check to see if rail voltages are on the speaker terminals!

Time for a beer to get my strength and carry the beast back to its home. Connected with Quicksilver Gold on the input connects at the back of the amp (other ends and speaker connects can wait for another day and preamp work).

Listening:
Powered up and immediately things sounded very liquid and clear. I will not go into all of the superlatives, but I will say that sound stage, depth, low level ambiance are very very good. And the bass is much better formed and dynamic. That certainly speaks to the old PS and rail caps not been able to deliver the goods which led to distortion from the amp modulating its own rail voltages... which says that the amp now has greatly improved power supply rejection. The above comments are from 3 hours of listening. I think that things were fully in focus after 40 minutes... at least for now.

So the Blackgates seemed to have reached a comfort zone rather quickly. I know that QS and QS Gold take a while to settle down, and I will have to wait on that. But most of it is in my PS and only signal path affected is the input WBT RCA jacks and plugs.

I am very enthusiastic about this result.

Prior mods: Some of the .47uF caps were on the PSCB from before and that made a big improvement at the time*. The WBT jacks were on from before and the speaker leads were changed before from lugs, washers and nuts to soldered onto the output jack posts. *(Perhaps the main PS caps are not very good and the source of some of this improvement.)

Other observations:
The output protection CB looks for DC offset and other problems and that controls the output relays. Each channel has a massive relay with four heavy contact points per channel. With the DC coupled output, there is a servo circuit to maintain a zero DC offset voltage. Otherwise... I do not like speaker relays and have defeated these in the past. But perhaps this amp is otherwise a bit rough starting up.
The speaker ground return current goes through the output CBs. I would prefer that this current went through a different return wire, and that the on board rail caps had their own returns as well. I have experimented with these things before and found them to be worthwhile. But doing things like this on this amp would be a major job. Note that the feedback 23K-1K resistors refer to hard this ground reference, and that speaker currents passing through the single ground wire can modulate the ground reference for the feedback.. aka distortion.
Fuse clips are very stiff, but not magnetic as I had presumed.

I would recommend these changes, on this or similar work on any amp. I would not recommend anyone inexperienced working on the 8008 output boards, but the PS board is easy. It is simply a difficult amp to do relatively simple output board mods.

I think that my first experience with Blackgate has been very positive. Certainly has lived up to all the things said previously.

I can supply circuit diagrams for the output, PS and balanced input CBs. Do not have anything for the protection CB. Can supply photos as well.

System: Denon 2700 with major analog output topology changes, and 'K' level DACs, Conrad Johnson PV-5 tube preamp heavily modified, Teflon interconnects and WBT jacks and plugs, Magi III's with upgraded PP caps and wire-connector improvements, 2xDD15 subs. (Those subs really made the changes in bass from the amp very apparent. I drive the DD-15's from the amp speaker posts as I do not have a buffered output on the Pv-5, so the subs listen to what the amp is doing.)