NAD 3020A Bias

[Corona Blue]

Greetings all-

New to the forum, first post. Very happy to have found this place.

I picked up a NAD 3020A recently and have been going through it. The unit was working, but it clearly needed new e'lytics. I replaced all with Nichicon KZ, FG, or ES "Muse" series caps and changed the old 1mfd coupling caps between pre/power with 1mfd Roederstein film caps. Also while in there, changed the .68mfd caps in the EQ network with Roederstein film units. All of this has made an audible improvement in the unit's sound, however, it does seem to have a bit of distortion at lower levels that I would like to clean up. I did put a meter across the 1ohm resistors (R653 and R654) and observed 0.00mV at warm idle. The service manual indicates a window of 30-60mV, so I think I may be getting close to finding the problem.

I realize that these amps were never the cleanest sounding, but just thought I would ask the experienced NAD techs here if there is perhaps a "sweet spot" within the specified window to shoot for when setting the bias on these? I planned on installing a couple of .5 watt 1K ohm cermet pots to replace the biasing resistors (RX1, RX2) just to make the job a bit easier.

Thanks for any help on this, and for the great forum!

Allen

[Ian Finch]

Assuming you already used the search button to check for existing threads on 3020/3020A problems (there have been hundreds) you probably found the 3020 schematic and range of 5-11 mV set by VR6 in the original. Here's an example: Help NAD 7225PE (3020) receiver output transistors replacement.

The 3020 design generally (there are several very similar models) has a crude but effective O/P stage stabilisation scheme that will be compromised by tweaking bias, IMO.

Read as much as you can before deciding to risk making such adjustments. Ive seen many binned 3020 type amps for just that simple reason.

[Ian Finch]

Sorry, no adjustment on reflection..VR6 is offset. Other comments still apply.

[Corona Blue]

Thank you, Ian!

Yes, I plowed around searching for 3020 repair threads and such before posting here. I didn't find anything specific to the biasing of the 3020A, other than the 3020 service manual itself.

I do appreciate your caveat, and noted that other threads made mention of the inherent limitations on these units. Part of the romance, I suppose...

I really only intend to bring the unit back into spec after renewing the e'lytics, hoping to smooth out it out a bit.

Best,

Allen

[Corona Blue]

Well, just wanted to follow up with this thread: Indeed, as Ian suggested, the idle current adjustment at resistors RX-1 and RX-2 sounded best as set from the factory.

It was a bit puzzling to me that the factory installed biasing resistor values yielded an idle current of about 35ma on the left and about 55ma on the right. What seemed proper was to match the idle current from channel to channel. I tried several different idle current settings, within a generally safe zone as stipulated by the service manual, but nothing sounded any better. And again, as Ian suggested, I found also that the power regulation on these units is a bit feeble as the voltage would increase on one side as the value of the adjusting resistor on the opposite side was increased and vice-versa. Thinking it all the way through and looking at the schematic, it makes perfect sense. I tried matching the current of each respective output pair at around 40-45ma, but when listening to this found it to sound dull and veiled. In fact the right channel was lower in volume by around 2 to 3 (just enough to be clearly noticeable) and the amp lost what dynamic character it once had. I put the original biasing resistors back into their proper place, and it sounded balanced and dynamic again.

So, the question I have is, when building one's own SS push pull amp, how does one determine the biasing requirements of each respective transistor pair? Obviously, in the case of this old 3020A (actually it's a "B" based on the schematics I have) the optimal current requirements for each pair was quite different. What methods are normally used to arrive at the optimum current for each push pull pair?

The unit sounds very sweet now as long as the volume is below the mid point. After that I think the signal begins to squish too much and the articulation is blurred. But with some good speakers, there isn't much need to turn the knob much past the 1/3 point... plenty loud!

I did change the ceramic caps in the driver network using 500V silver micas...



I'm still undecided if I like it better this way. On one hand, it does sound a bit more detailed and open. But it seems to border on being a bit sharp. Anybody here prefer using the lower voltage ceramics for this kind of network application?

Sorry for all the questions, but thanks for the good information!

Best,

Allen

[djk]

Try adding some 4R7 1/2W base resistors to the outputs.

[jaycee]

This amp was made when the 2955/3055 was produced using the "Hometaxial" fabrication process - which is why there are no emitter resistors. If these have been replaced by modern 2955/3055 transistors of the "Epitaxial" type, then emitter resistors should be retrofitted. 0.33 ohm 3W resistors can be fitted underneath the PCB I believe. Not doing this will lead to the amp going into thermal runaway. This omission is also responsible for the bias being a PITA to set.

[Ian Finch]

Originally Posted by jaycee

This amp was made when the 2955/3055 was produced using the "Hometaxial" fabrication process - which is why there are no emitter resistors. If these have been replaced by modern 2955/3055 transistors of the "Epitaxial" type, then emitter resistors should be retrofitted. 0.33 ohm 3W resistors can be fitted underneath the PCB I believe. Not doing this will lead to the amp going into thermal runaway. This omission is also responsible for the bias being a PITA to set.

Now that's interesting, Jaycee. It might explain why I saw a few 3020s with more failures after repairs, even by pros. I would not put the date of changeover from homotaxials that late though. 3020s were fitted with any old 2955/3055s, National, Fairchild, CSF Thomson, whatever. We did not see them until around 1980, long after phasing in of Epitaxials, I would have thought but perhaps different manufacturers retained the earlier process whilst others changed asap?