All sounded and looked good to me! Especially mounting those diodes off the PCB. Typical and a shame caps are mounted so close to hot transistors and diodes, etc.

I was thinking there must be another fault because the video glosses over any current calculations you did. I'm glad to see you got to the root of the problem, and sad to see that NAD gave so little regard to where all that heat would wind up. You can't PCB mount components that get their legs hot enough to melt solder, so increasing the surface area of your parts was the only way to go.

went to check the schematic myself, and it would have been much simpler if they added a couple of outputs in the power transformer with lower voltage instead of lowering the voltage from + and - 55 to + and - 26, the only question I have is how long the amp lasted until it showed the problem.

Hi sir, where I can find a Galvo mirror for 1064nm ? Do you know some seller? Thanks in Advance.

I wonder if they knew who you are and your background, then would they still be making those comments...? I'm loving your videos! Thank you for taking the time to produce them!!

Yeah - you suspect there is some other fault. But I trust the skills of our fearless leader.

​ @mister kluge I suspect I know full well the reason why they did it that way (since I am not the designer of the product I cant know for sure, and I would love to be proven wrong), and it is much, much more mundane than many would think... My guess is that the whole regulation section original design was done for an older/other model, and then carried through to the new one (nothing wrong with that usually) however the main PSU for this specific design runs a higher voltage than the one originally meant for this design. Instead of reworking the design for the section to take the higher voltage, they just bodged in a couple of zeners here and there to drop enough voltage to get into the limits. Again I guess that those supplies are only for the preamp/driver sections, miscelaneous analog and maybe logic, thus not to terribly high current, so this design was deemed acceptable. In the long run however this leads to failure either due to thermal cycling or just overload.

@@pelor92 I think you are on to something. They had this custom made, expensive toroidal transformer with the 55V extra secondary from a previous model. Mr. Pinch-Penny working overtime in the procurement department refused to shell out the outrageous NRE for the correct new transformer. I bet you there was an interesting meeting at the company and the engineers were not very pleased with the outcome. So they used what they had and shed the extra 20V in resistors and zeners and PCB cooking heat while cursing accounting. Making this all up, but I sure have lived it in my engineering life. And then made a couple such poor decisions myself later on as a manager. I can't fathom any other reason why you'd use a 55V secondary for a 26V linear supply.

@@CuriousMarc "I can't fathom any other reason why you'd use a 55V secondary for a 26V linear supply." - This kind of thing used to be *very, very* common in integrated amps 30+ years ago. Just cost-cutting, especially when the preamp section is little more than an RIAA preamp. Who'll care about 2 W more power dissipation when the power amp dissipates an order of magnitude more than that? Plus I guess that transformers with extra secondaries are expensive special orders (especially on smaller runs) with little chance of second-sourcing. NAD neither have the kind of production volumes of Yamaha or Pioneer nor are they a premium brand with massive margins, so they really have to watch their BOM cost. It shows. Their stuff is well-engineered but parts quality is really rather _meh_ at times.

@@CuriousMarc Maybe a "fix" would be to add one more transformer with a lower secondary voltage inside the case somewhere, and use that instead as supply?

My 3020 is still running sweet. All original parts.

@@sarkybugger5009 I'll admit my computer speakers are driven by a newfangled piece of electronics - a D3020. Apparently they wanted it to fill the same kind of "not very expensive, not ultra-HiFi, but practical and sounds good" niche, and named it in memory of the original 3020. (It is nice, too - convenient set of inputs, will drive a reasonable selection of speakers, sounds good. I have some tiny Dali bookshelf speakers and it's all very pleasant.)

My 3020 is still working great too! I bought it second hand in 1996 :)

good 'un Is there any brand of mid-fi amplifier that you *would* recommend?

there are medical doctors, here there are electronic doctors. :-)

110C thermal fuse, or a 5W TVS diode across the main power fails might be good, along with a soldered in wire ended Littelfuse fuse. 5W 30V device by the pass transistor collector and a 0.5A fuse would provide good protection.

FS means fusible. The resistors were replaced with fusible ones when available ( could not find the 51 ohm in stock unfortunately). There are also several crowbars in the design, which is generally well protected, and an over temperature protection circuit.

@darren murphy Why would you say that? I did check. The current draw on either rail of the supply is normal (237 mA for the +55V side and 170 mA for the negative side). Which means all the components were running within their power spec and the amp is functioning normally.

@@AstrosElectronicsLab I measured the current going through the components on either rails of the supply. The current is normal (237 mA for the +55V side and 170 mA for the negative side). So all the components were running within their original power spec and the amp is functioning normally. The larger components and heatsinks are used to decrease the operating temperature so they don't degrade the PCB, which is what failed, not to cover any fault.

Oh really? Thanks, I did not know!

The other commonwealth!

Yes for once I agree with the electrolytic change, the one near the zener is probably a little cooked by now.

@@CuriousMarc Yes they are JRC, and probably are also Just Really cooked. I typically add a heatsink to high power diodes, using some copper shimstock cut into strips, with a hole drilled into it just to clear the lead diameter, and soldered right by the diode body, so the heat through the leads is conducted away. The major path for heat to escape those diodes is via the legs, not via the plastic body, so the tags on the legs is the best method. Saw this years ago in a Motorola diode application note for the high power Shottky rectifiers, and it really does work well, as they had a nice neat set of graphs showing the effect of various lead lengths, copper plate sizes and versus clamping the diode body in a copper strap attached to a much larger plate. going to make the mounting a little harder, but they do run 40C cooler with only a modest fin each side. As to the NAD branded capacitors, all they are is a generic JRC capacitor that NAD decided to buy a batch with a special sleeve shrunk on, in place of the original, or they just resleeved the cheap ones as a marketing point. Kind of like you can buy various forms of ****fire lithium battery, all with just recycled mystery laptop cells that have been recovered from ewaste, and all with 6-10Ah capacity marked on the 18650 cell, irrespective of what it was.

maybe that's why NAD didn't catch this flaw at the design and manufacturing stage if it takes a while for these amps to fail? cooked electrolyte and possibly general aging causing more leakage and increased power draw through those undersized components?

@@SAerror1 No, just the thermal cycles will degrade the diodes, and the resistor as well. the degradation is strictly a time and cycle thing, you need to run the resistor hot for a few hundred hours for it to degrade at the full rated power, and the same for the diodes. They just ran the parts at the full rated load, figuring the lifetime would be around 10k hours of operation before failing, and that the fusible resistor would fail first if the diodes went short circuit.

@@CuriousMarc That's what I thought when you installed this little radiator (the heatsink) right next to the (probably already cooked) cap there. Could be a good idea to replace the cap(s) and move it a bit away from the heat source.

Allmentux - Great idea! I second that.

Yep, something downstream is surely leaking, I’d be looking at those electrolytics or an IC that’s misbehaving, can’t imagine that circuit should be drawing more than a few milliamps with that amount of drop... agree, even with a hundred times the surface area the temps for that tiny original part are far far beyond anything reasonable and would have less than an hour to live for sure... the temperature of the larger parts will be much lower for a given power that’s exactly why they are physically larger, they radiate over a larger surface area, the thermal conductivity is much higher.

The resistors were still alive and well when removed them. Just burned the cheap PCB over time. No other hot component that would indicate a short downstream. Just a ballast doing its thing trying to drop 20V at a couple 100 mA...

CuriousMarc just saying 60c is at the absolute maximum rating for a 5 watt resistor and likely 20 times beyond a 1 watt... if nothing else is wrong then it’s about the worst design in history.... difficult to believe a component would survive long when stressed nearly 20 times its absolute maximum rating... if it’s dropping 20v with nearly an amp through it then a 25 watt would be the minimum power level for a resistor, seems rather insane... maybe incorrect but it just screams to me the current draw should be a couple tens of milliamperes not nearly an amp...

@@robertw1871 Resistors running at 60C are not 20 times above their ratings. Most resistors are resistors rated 150C, power ones will do more, up to 270C. Why does 20V equate to 25 Watt?

@@CuriousMarc Ohms law, you stated it has a 20v drop across it and nearly 1 amp of flow, that puts the power dissipation at close to 20 watts, if it's a 1 watt resistor then its 20 times its rated capacity or nearly that...

Good electro-sleuthing!

I get that feeling too.

I’m sure Louis would have something PP bus related to say on the matter... But yes I agree, he’s looking the symptoms not fixing it fully

I checked and the current is perfectly normal (237 mA in the + rail and 170 mA in the - rail). No excessive power consumption. The original parts were running within their rated power. That said they generated too much heat for the PCB which degraded over time. The bigger replacement parts are just easier to cool, not there to patch over anything.

Cooking them tends to speed that along... An engineered failure IMHO

A high ESR in a cap will cause that cap to heat, not the ballast resistor nor the zener. A shorted component would, but it would be very hot and obvious, and I did not find any. The fusible resistors were working fine when I took them out, they had not triggered. One zener had failed, but the other one apparently took the load and was still working. It’s mostly the PCB that did not like the heat.

​@@CuriousMarc Just check ESR and capacitance on caps around heatsinks and zeners they are just 2-15 Ohm resistors now with nanofarad capacitance.

I assume the issue is more of leakage (which is another way electrolytic caps degrade) but also some of these consumer gear are engineered within a hairs breadth of the limits in terms of thermal design

@dimmog Agreed. I think you have a high chance on being proven right (you mean 10 ohms ESR, not resistors - they’d fail open).

I thought that's where this was going. I think perhaps the galvos weren't up to the frequencies

Operative word is "WERE", they are now very low capacitance glass capacitors, with a breakdown voltage of around 80-200V

Well said never thought of it that way we’ll if I never need an emergency cap I can just look in my hot running amp

Was considering T778 and wonder if reliability has improved on newer models. Will investigate.

Replacing electrolytic capacitors, what a revolutionary idea!

Agree. The series zener diodes is a very poor design choice. I always avoided using zeners at all, actually.

All my friends who had NAD equipment rave (rightly) about the sound quality. The machines don't last that long though. NAD looks like a holiday love, beautiful but temporarily.

Marcus Wilson any brands you swear by?

@@chevon5707 Not Marcus, but I'd recommend Cambridge Audio.

Frickin laser

Seems like a missed opportunity for some oscilloscope music....

I'd suspect some other fault that's loading those rails down too much - maybe bias voltages out of spec in the power amp section or something? To me to replace the parts with 5x higher wattage and they still get too hot would indicate some other problem somewhere. I mean the design couldn't be that bad to begin with, could it?

Do you feel amps really have a "sound"? A good amp should only faithfully reproduce the input signal adding nothing but gain?

Combo of amps and speakers definitely do, some very markedly so.

@@CuriousMarc I got it paired up with some Bowers & Wilkins DM309's, so I can completely agree with you there!

@@larryl730 Absolutely an amp can have a different sound to another one. Its not always about just faithfully reproducing the input they all add their own character to the sound. For example, I find most mainstream amps like Yamaha's and Sony's produce a very bright sound as opposed to NAD's warmer sound. These nuances are even more pronounced with different amplifier technology like valve amps.

@Deano023 That was also my first NAD amp. Agree with you, the more refined top end is what drew me to the brand (plus I could afford it!).

The resistors didn’t burn. They were perfectly functional when I took them out. One of the two zeners had failed, which the other one took in strides, it was still working too. The ballast circuit just ran hot for many years and damaged the very cheap PCB over time. Something we often see in circuits that uses ballast resistors in our vintage equipment (just happened on our ASR33), but also sometimes in modern stuff.

@@CuriousMarc Thanks for the clarification. Love the heat sinks. Never can be too cool, I suppose. Cheers.

Yes I did. Everything appears to be running as designed. Resistors were still working. However one Zener had failed, no doubt from thermal stress. But even that was not enough for the other zener to fail or the fusible resistor to let go. I it just damaged the cheap PCB over time, and the resistors desoldered themselves over thermal cycling. Bad thermal design.

It’s been problem free since the repair.