Restoring a NAD 3020B amp.

Hello and welcome to WhatIsHiFi blog which I hope you will explore and enjoy.

Everything started back in 1986. After having all-in-one systems from KOYO, Crown & Sharp and realizing as a teenager that I am in love with Hi-Fi, time had come for my first separates system. After some research and realizing the funding limitation I had to face, a NAD 3020B amp plus a NAD 6130 tape deck plus a NAD 5120 turntable was bought as a Christmas present. I still remember the thrills I got listening to Dire Straits / Private Investigation through this system at a very nice and proper demo room at Hellas Sound, the ex NAD distributor in Greece back then, offering a unique 5 years warranty. I also remember that I wanted a Dual 505-2 turntable, but didn't have the money.

Anyway, at some time where we thought that DAT and then CD-R is the future, the (malfunctioning) deck was given away. But the amp and the turntable is still with me, so with the encouragement and valuable help of my friend Panos, why not restoring them.

This little amp has become somehow legendary from the moment the first 3020 was born. I remember very clearly that it had nothing irritating or insulting at it's sound. It helped you listening and listening without fatigue. The one I have is just 35 Watt RMS and as a teenager I remember I wanted more (of course!), however these 35 watts were very real. When the amp was new, the only comparison I had made keeping everything else the same was with a borrowed Luxman integrated; it must have been an L-220 or an L-230. I felt that with the Luxman, the low end was stronger. When I purchased a cheap Sansui B-1000 power amp and used the NAD solely as a pre amp, I remember that although I got the power boost I needed, the magic was somehow gone and the saxophone at Whitney Houston's "All the man I need" started sounding harsh. I then got a Luxman M-03B as a power amp, after a while also got the matching C-03B pre amp and since then the NAD was given to my father, which means it was practically left alone unused.

A lot of dust in my NAD; even though it was stored indoors, looks like it was abandoned outside.

Anyway, let's start with the repairs. The amp had become unlistenable, with many pops and clicks as well as loosing sound at one channel. This is why there couldn't be even a basic before / after kind of listening comparison. Since the amp was quite old and built on a budget, we decided to go for a full electrolytic re-cap, something that proved wise because all the TEAPO capacitors it had might have not leaked, but they where way out of specs; some of them with ESR of 17Ω! 

At this point, let me make some comments regarding caps. We have found caps at 90's amps to be OK (Luxman M-03). We have found that the Nichicon power capacitors at a 1983 Sansui Z-9000 receiver were totally out of specs, while Nichicon signal capacitors at the same receiver were pretty much OK, defying aging. Eventually, an electrolytic capacitor will need replacement but we don't know when or where. Also, we don't know the effect an old capacitor has at the sound. To make a long story short, don't get hysterical regarding capacitors. Just keep in mind that capacitors involved in the power supply need to be healthy, otherwise you might jeopardize your equipment.

Also, changing capacitors will most probably alter the character of the sound of the equipment. I was discussing this with a friend who is at the HiFi and speaker repair business for some decades now (and he is 100% down to earth with HiFi), and he told me that back in the 90's when he had a Pioneer A-07 amp, he got the best super duper Nippon Chemi-Con for audio capacitors he could find and replaced the original 4 big power capacitors that this amp had. After that, the amp was sounding a bit on the bright side. Thinking that this might be just his idea, he installed the original capacitors back; this way he confirmed that at that system he had, the sound was better with the original capacitors, event though they were "inferior". Please mind that I wouldn't expect the sound to change by changing the capacitors at this particular position but it did; so, think what happens when you change caps at the signal path. So, once more, don't get hysterical regarding capacitors. Ideally we would need to try various capacitors at various positions at an amp one by one and see what happens but this is not very practical.

As always, one of us would see the actual capacitor on the board and with the aid of magnifying glasses would read the values (including temperature), put a mark on the top of the cap (so we know with which caps we are "done") and the other would search at the schematics what is the role of the particular capacitor and order the optimum replacement. Since sometimes the replacement capacitor is not an exact match but a better one, or there are multiple types of caps with same values, you better keep notes of  what you actually order and the position you plan installing it; at this project we didn't and it took more time to find out what goes where than actually replacing the cap. If you order from Mouser, you can include the position of each capacitor in your order and they will come in separate bags with this position printed on the label. Other trusty on line sources for capacitors (except your local store) are TME and Digi-Key.

After deciding and getting done about the replacement capacitor, we have put a mark on the actual capacitor; black for the power caps, red for the signal caps (we marked more caps as red after this photo was taken).

On choosing the capacitors you will use, I suggest you to go for a good brand like Nichicon, Kemet, Rubycon, Elna, Vishay or Nippon Chemi-Con. For the ones used in power supply rails, you need low ESR models. For ones used in power supply rails but directly related with the sound (like the big caps that feed the power amp section), it is logical to go for audio capacitors. For caps in the signal path, you better use bipolar / MUSE electrolytic caps or film / WIMA caps - the film ones have 10% tolerance compared to 20% of electrolytic. If you go for different type / higher voltage / higher temperature capacitors (the benefit of generally installing a higher voltage or a higher temperature capacitor at positions that get heated, is that it will almost certainly live longer), make sure they will physically fit; however, if you are replacing old caps, it is nice to know that new caps are much smaller in size. Same thing is to be considered regarding increasing the capacitance of caps used in the power supply, plus understanding if the bridge rectifier is capable to deal with the increased power demands of initially charging the beefier caps when you turn on your equipment.

Look how smaller in size capacitors have become; the one on the right has bigger capacitance and still is smaller in size.

Sometimes, you will have to make decisions (actually decide what is the best compromise) based on availability and physical space available; for example, choose between audio caps 4.700 μF or low impedance 5.600 μF.

Before recapping a NAD 3020B, keep in mind that it was a budget amp and this reflects also to the engineering. We found the schematics disagreeing with the capacitor list which again disagreed with what we actually have found inside the amp. 

There were nowhere in this NAD bi-polar capacitors from factory, which we strongly recommend to use for the signal path.

Last but not least, we found some 105°C capacitors but they did not seem to have some logic behind them - anyway, using higher temp capacitors is only good regarding the durability of the cap. I suspect that there might was a shortage of 85°C capacitors at production, so they would go on with 105°C. Also, there was a 100V capacitor with no apparent need to be 100V. However, increasing the °C or the V at a capacitor, almost certainly will lead to a longer life.

Here are the actual factory caps that we took out:

- 2x 4700μf, 50V, 105°C.

- 2x 2200μf, 6.3V, 105°C.

- 2x 1000μf, 10V, 105°C.

- 2x 330μf, 50V, 105°C.

- 1x 47μf, 100V, 105°C.

- 2x 220μf, 10V, 85°C.

- 3x 47μf, 50V, 85°C.

- 2x 47μf, 35V, 85°C.

- 2x 100μf, 10V, 85°C.

- 2x 47μf, 35V, 85°C.

- 6x 47μf, 25V, 85°C.

- 2x 33μf, 25V, 85°C.

- 4x 47μf, 10V, 85°C.

- 4x 22μf, 10V, 85°C.

- 2x 10μf, 16V, 85°C.

- 4x 4.7μf, 50V, 85°C

- 1x 1μf, 50V, 85°C.

- 2x 47μf, 10V, 85°C.

- 2x 4.7μf, 50V, 85°C.

- 4x 1μf, 50V, 85°C.

- 2x 0.68μf, 50V, 85°C.

A nice bunch of old capacitors.

Since we decided to slightly increase some capacitance in the main power supply and in order to be at the safe side, we also changed the bridge rectifier with a beefier KBL606G.

So, here is what we choose to go for, all but two are Nichicon brand. Total cost for everything including the rectifier bridge was around 36€ (24% damn VAT included, transportation cost excluded).

C401 & 402, signal path, bipolar 4.7μF 50V (same values as original but better capacitor type).

C407 & 408, 2200μf, 6.3V, 105°C low ESR (same values as original).

C409 & 410, 33μf, 25V, low ESR (same values as original).

C417 & C418, 47μf, 25V, low ESR (same values as original).

C421 & C422, signal path, bipolar 10μf, 16V, (same values as original but better capacitor type).

C423 & C424, 220μf, 10V, low ESR (same values as original).

C513 & C514, 47μf, 10V, 105°C, low ESR (same values as original although we feel 105°C is not necessary).

C517 & C518, 47μf, 25V, low ESR (same values as original).

C519 & C520, signal path, bipolar 100μf, 16V, not 10V & 105°C as original (same capacitance as original, 16V because we couldn't find 10V bipolar, better capacitor type. We feel 105°C was not necessary in the first place. Keep in mind that there are no 105°C bipolar caps).

C529 & 530, signal path, film WIMA 0.68μF 50V (same values as original but better capacitor type, be aware that these positions were missing from the service manual capacitor list).

C525 & C526, 47μf, 10V, low ESR (same values as original, be aware that these positions were missing from the service manual capacitor list).

C531, 47μf, 100V, 105°C low ESR (same values as original). However, we believe that neither 105°C nor 100V are necessary here. CAUTION! In the service manual capacitor list, this position is wrongly mentioned as 4.7 μF instead of 47μF. At the schematics, this capacitor is mentioned as 63V instead of 100V our unit actually has. To make things even more confusing, the position name on the circuit board was covered with glue.

C532, 4.7μf, 50V, low ESR (same values as original).

C533, 47μf, 50V, 105°C low ESR (same values as original although we feel 105°C is not necessary).

C601, 602, 603 & 604, signal path, bipolar 1μF 50V (same values as original but better capacitor type).

C613 & 614, signal path, bipolar MUSE 47μF 25V (original was 10V but we couldn't find bipolar 10V so we choose 25V and a better capacitor type).

C621 & C622, 1200μf, 10V, 105°C, low ESR (original was 1000μF but we increased it, also we feel 105°C is not necessary).

C625 & C626, 47μf, 25V, low ESR (same values as original).

C701, 702, 703 & 704, 22μf, 10V, 105°C, low ESR (same values as original although we feel 105°C is not necessary).

C803 & C805 at the power indicator board, 4,7μf, 50V, low ESR (same values as original, however at the schematics one of them is indicated as 1μF, be aware that these positions were missing from the service manual capacitor list).

C901 & C902, 5600μf, 50V, 105°C, low ESR (these are the two main big capacitors, originals were 4700μF but we increased it).

C905 & C906, 330μf, 50V, low ESR (same values as original).

C907 & C908, 47μf, 50V, low ESR (same values as original).

C911 & C912, 4.7μf, 50V, low ESR (same values as original).

C913 & C914, 47μf, 35V, 105°C, low ESR (same values as original although we feel 105°C is not necessary).

In order to be able to disolder the capacitors, you obviously need to remove the top and bottom cover; it is so nice when you can remove the bottom cover of a Hi Fi, repairs are so much easier and I do not understand why all manufacturers don't keep it this way.

This metallic bridge must be removed in order to be able to fully recap the NAD.

You will also need to remove a metallic bridge from the bottom which is a bit tricky but not difficult. In order to do so, you will need to remove the plastic face plate (which till now I thought it was made of steel). You have to remove two screws that hold the face from the top and two from the bottom. You will also need to remove two screws that hold the power indicator led board.

The power indicator board can easily be removed in order to free the face plate and also replace these two little capacitors. But be extra careful when replacing them, the rails are very agile.

Now, here comes the tricky part: you will also need to remove bass, treble, balance and volume plastic knobs which in Taiwan 1985 they thought it was a good (good? great!) idea to also put some glue to hold them. This is so stupid and totally unnecessary and will stress the potentiometers but you have to do it. In my unit, the three knobs came out just by pulling by hand but one of them needed a cloth around it (in order not to scratch it) and a tool to pull it out. After doing so, you can pull out the face plate, the cord for the power switch / indicator on the left is long enough and you don't need to do anything.

Not a smart thing to do. Please remove this glue before you put back the knobs, you do not need it.

The metallic bridge we try to remove has one screw at the front, one at the back and two on the board. Do not mix the screws, they are 4 different screws and you have to remember where each goes. If you mess it up, the metallic screw is at the front, the taller of the 3 black screws is at the heat sink, the scratched one is at the back and the one remaining is at the board at the back.

Now you are free to remove all the capacitors. As a general rule, we keep an eye regarding where the negative leg actually is before removing a capacitor; you never know if there is a wrong indication at the board. In our unit there was not, everything was marked correctly.

With the aid of a flash light you see what you need to remove and you disolder it with the appropriate tool. You have to be fast and accurate in order not to stress the board. After removing a capacitor, you put the new one in the position always being careful regarding polarity and you move to the next one. After having replaced a bunch of them, go ahead and solder them.

With the help of a strong flashlight and by creating a shadow with a tool or your hand, you can locate the position of the capacitor you want to remove. Photo is from a Sansui restoration project.

Luckily, even though some caps where glued on the board, the glue that was used was not the notorious "glue of death" type that becomes conductive over time. We removed this glue and we didn't feel that we should glue any of the new capacitors; none was heavy enough to be a problem and according to our knowledge, this unit has no intentions immigrating.

After we replaced all the electrolytic capacitors, we cleaned all the RCA plugs; I have mentioned at this post how much of a difference this can make. You don't have to follow same procedure, but please do make some basic cleaning.

We took out the 4 fuses, in order to clean them and their contacts.

We also sprayed the appropriate cleaning / lubricating things at all the switches, potentiometers and headphone jack.

We treated accordingly every contact, potentiometer and switch.

Back in the 80's I remember that the purists were saying that the headphone out degraded the sound quality of this amp and found it so funny; however Panos discovered that this is true. At this amp, the signal for the speakers passes through the headphone jack; when you plug in your phones, the jack presses a lamina and this way the speakers are disconnected. We were tempted to by-pass the headphone out circuit but eventually we did not; still, it seems not a bad idea to do so.

Panos also saw that the "lab in" input is more direct, with only one capacitor at the signal path instead of two capacitors that the "normal in" has.

Also, for the ones of you wondering about what the soft clipping switch does, it is handy when you wanna party; due to the limited power this amp has, most probably you will cause clipping which can burn your tweeters. So, in those cases put this switch to on, the clipping will come progressively and will be more kind at the beginning. For normal listening, leave this switch at off.

The speaker matching switch is what it says, put it at the 8Ω position for speakers with nominal impedance of 6Ω or above and at the 4Ω position for speakers that are considered as "difficult" loads.

The NAD 3020B nice and clean after the restoration. Please do note that the two green bipolar MUSE caps at the front middle (behind the balance pot) are unnecessary; we just didn't have common type available at that moment. I haven't put them on the list as bipolar either.

After finishing everything, since the unit was working before the restoration and we were very careful to do everything right, we decided to turn it on without the aid of a variac and a lamp. We only checked from the very beginning that the transistors were not heating excessively. We also checked DC offset which was found to be within specs. Bias adjustment at this model is painfully complicated, but on the other hand this has the positive effect that it is designed to stay steady over time; we suspect that it will only need adjustment in the unfortunately event you need to install new transistors.

After establishing that there was no explosion or magic smoke, we connected our stuntman speakers; everything was sounding normal. Except when we turned the unit off, a rather loud click was heard through the right channel; something that would not happen if instead of pressing the power button, we just unplug the amp from the mains. This made us suspect the power switch, however facts were misleading. You see, if you don't turn off NAD 3020B from the power button and just unplug / plug it in to the mains, the primitive muting cycle that happens every time you turn on / off the amp by the power button does not occur. Before understanding that, we took off the power button, made a bath with WD40 and changed the caps that had on it - of course as you guessed nothing changed. We also inspected / cleaned the joints of Q509, Q510, R549 & R510; these are the components involved at this mysterious muting cycle / protection this amp has. Things became a bit better but I suspect it had nothing to do with what we performed. To make a long story short, this click noise most probably comes from the Q509 & Q510 FETs that are used as solid state switches here. Since the noise was eventually no threat, we called it a day and proceeded with the listening test. However, if you want to improve a NAD 3020B you might try (at your risk) removing Q509, Q 510, R545 & R546; they are all in the signal path.

After leaving the amp to play for some hours and after finding  the correct polarity of the amp, we connected it to Panos main floor-standing speakers and started listening. We used reference Jazz recordings, Dire Straits, even Robert Miles. The sound we got was such a pleasant surprise. Everything was sounding good and nice. Nothing was disturbing. This amp is wonderful in encouraging you to just enjoy the music and forget about judging the sound.

The NAD ready to be tested. As a source, we used our trustworthy tweaked Pioneer CD player.

Even the much demanding in the low end "Willow Weep For Me" from album "Greensleeves on gold" was reproduced with the usual punch; the windows, the chairs and our stomachs felt it. At that point, I wandered how loud was I listening 3 decades back. So here it goes away the "somehow weak in the low end" I remember; however, with some recordings, I felt that you had to crank up a bit the volume - but always within a civilized range - in order for the amp to perform better in the base.

Having combined experience with Yamaha, Sansui, Luxman, Parasound, Rotel, Pioneer, NAD and custom made pre / power / integrated amps, me and Panos realized that just a humble amp like this NAD is capable to make you happy in listening your favorite music. Yes, compared to the monster Parasound and Luxman amps we currently have, the soundstage with the NAD is a bit less three dimensional and airy. Yes, some very fine details are not served to the listener so much as self explanatory. But you know what? Everything happens in a nice manner. You just technically happen to know that a little something is missing, because you have a high revealing system. Still, you are not suffering. You are not left there wanting for more, you just want to keep listening to the music.

So, I have to say that if this amp is properly recapped it is recommended. Is it the best for the money? I can't know, I haven't owned every single amp that is out there. But I have to say that it encourages you to keep listening. It has MM/MC phono stage. It has loudness button, for those late night low volume listening. It has pre out / power in, meaning you can gradually upgrade to pre / power stuff. You might miss a remote control but you will never miss the desire to listen to your favorite music. Is it a giant killer? No, but this was not what it supposed to be; it should help you love music and HiFi, which in every aspect it succeeds doing. You worry less for the machine and you enjoy more the music.

Happy listening as always!

Many thanx to Panos.

Christos

PS. Have you ever noticed that if you unplug the speakers from the NAD, crank up the volume and put your ear on the top cover, you can hear the music coming from the inside?