My digi-key package arrived today 
First, I put my 3020 back in the system for a quick listen of a few well known tracks, including one where it falls apart badly. (just about anything on Robbie Robertson's 1st album)
Then, I upgraded the 3020 PSU caps from 2200uF to 5600uF, with very slightly larger diameter units from United Chemicon SMG series.
All went smoothly, and my solder job even looks better than original (which isn't saying much...).
Back to the living room, the 3020 sounds much more composed on Robbie Robertson, where previously the bass was out of control and flabby - now it is much better controlled, although still a little heavy handed and loose for my tastes.
Now I have most of the caps to replace on the 3020, but really wanted to try the PSU caps on my 3125, which has been my main amp for 20 years now. (And was just brought back to life a few weeks back, when I decided to bypass the infrasonic filter section in the pre-amp, which was slowly dying.)
I had ordered 10,000 uF Panaonic TS series, to replace the stock 4700uF units in the 3125. They were snap-ins, so install was quick and painless. The old units were ELNA, and looked ok, except for slight bulge in one.
Listening to the 3125 didn't reveal as much improvement - it already sounded much cleaner and more dynamic than the 3020, especially on complex tracks, when the 3020 seems to just lose its composure. There is a bit more punch on very dynamic tracks, but the lower bass on the 3125 is not nearly as heavy as the 3020, which still seems too boomy to me.
One more swap for tonight - before putting the 3125 cover back, I decided to replace the 1000uF reg PS caps, this time with same size Panasonic KY, which are much smaller diameter, but same height. I have heard the smaller caps should remain the same size, including on the reg PS. Here, I had a problem with one solder track on the board coming loose, but was able to bend the lead back over a secure section and make it work. More care next time...
On listening, there seemed more of a difference now than after replacing the big caps - detail in the mids and highs seemed better, although bass seems about the same.
I'm still a little puzzled about the dramatic differences in the bass between the two amps. Next I plan to replace most of the remaining small caps on the 3020, in hopes of bringing back some clarity and definition. Then maybe I'll try a chipamp with the 3020 as a pre-amp.
First, I put my 3020 back in the system for a quick listen of a few well known tracks, including one where it falls apart badly. (just about anything on Robbie Robertson's 1st album)
Then, I upgraded the 3020 PSU caps from 2200uF to 5600uF, with very slightly larger diameter units from United Chemicon SMG series.
All went smoothly, and my solder job even looks better than original (which isn't saying much...).
Back to the living room, the 3020 sounds much more composed on Robbie Robertson, where previously the bass was out of control and flabby - now it is much better controlled, although still a little heavy handed and loose for my tastes.
Now I have most of the caps to replace on the 3020, but really wanted to try the PSU caps on my 3125, which has been my main amp for 20 years now. (And was just brought back to life a few weeks back, when I decided to bypass the infrasonic filter section in the pre-amp, which was slowly dying.)
I had ordered 10,000 uF Panaonic TS series, to replace the stock 4700uF units in the 3125. They were snap-ins, so install was quick and painless. The old units were ELNA, and looked ok, except for slight bulge in one.
Listening to the 3125 didn't reveal as much improvement - it already sounded much cleaner and more dynamic than the 3020, especially on complex tracks, when the 3020 seems to just lose its composure. There is a bit more punch on very dynamic tracks, but the lower bass on the 3125 is not nearly as heavy as the 3020, which still seems too boomy to me.
One more swap for tonight - before putting the 3125 cover back, I decided to replace the 1000uF reg PS caps, this time with same size Panasonic KY, which are much smaller diameter, but same height. I have heard the smaller caps should remain the same size, including on the reg PS. Here, I had a problem with one solder track on the board coming loose, but was able to bend the lead back over a secure section and make it work. More care next time...
On listening, there seemed more of a difference now than after replacing the big caps - detail in the mids and highs seemed better, although bass seems about the same.
I'm still a little puzzled about the dramatic differences in the bass between the two amps. Next I plan to replace most of the remaining small caps on the 3020, in hopes of bringing back some clarity and definition. Then maybe I'll try a chipamp with the 3020 as a pre-amp.
Hi,
the 3020 was designed for cheaper systems and these usually had fairly small speakers that had poor low end extension.
NAD may have engineered in a small amount of bass lift to augment the falling bass of the partnering speakers.
In those days, at least in the UK, there were a band of good audio retailers who compiled systems that made the best from matched components (from different manufacturers) that when swapped around may not perform as well.
the 3020 was designed for cheaper systems and these usually had fairly small speakers that had poor low end extension.
NAD may have engineered in a small amount of bass lift to augment the falling bass of the partnering speakers.
In those days, at least in the UK, there were a band of good audio retailers who compiled systems that made the best from matched components (from different manufacturers) that when swapped around may not perform as well.
AndrewT said:
Wouldn't this sort of thing show up on freq resp measurements? Or is it more of a dynamic issue with how the amp interacts with the speaker load?
I've now spent some time studying the mods listed here and plan to try removing the tone and balance controls next. I've never used them as I like a flat response, and my speakers are very flat also.
Perhaps cleaning up the circuit will affect the bass response; we'll see.
Right now I'm listening to my 3125 again after replacing a few more caps in the power amp section based on what I had on hand. (the largest feedback caps are still original). This seems to have cleared up the previously distant sounding treble and upper midrange, but there is a certain harshness in female vocals now that wasn't there before, especially as the volume increases above low levels.
The original caps in the 3125 are ELNA, and I'm not sure what the 3020 are - the symbol is a triangle with little squares at each corner. It would be interesting to measure the old ones as they come out to see how far off they are...
I'll report back after I do the tone control mods...
Hi,
Nad 3020 does not have bassbost in form of some EQ.
The amp has many sources for resistance to be added from power transistors emitter junction to speaker output, note this is outside feedback loop.
Signal has to pass: polyswitch - thin cables to headphone jack - over the headphone jack bypass - trough thin cables again and finally to speaker output. This adds series resistance and I have measured it to approx 0.3Ohm, most of it over the polyswitch (beware).
Speakers will of course be more or less affected by this as Q will rise.
Nad 3020 does not have bassbost in form of some EQ.
The amp has many sources for resistance to be added from power transistors emitter junction to speaker output, note this is outside feedback loop.
Signal has to pass: polyswitch - thin cables to headphone jack - over the headphone jack bypass - trough thin cables again and finally to speaker output. This adds series resistance and I have measured it to approx 0.3Ohm, most of it over the polyswitch (beware).
Speakers will of course be more or less affected by this as Q will rise.
TDWesty said:
I did the step 1 mods (remove tone controls) at the peak audio site above, but only on the right channel so far. I still need to get some better resistors to do it properly. So far the amp stills works
but the gain reduction is evident. I'm not sure of the reason for reducing the preamp gain from 6 to 3. I did notice that there is a typo in this section - he refers to "F" connections on the PCB corresponding to circled B on the schematic for the right channel. On my amp, it is E for right channel (circled B) and F for left channel (circled A). I spent a lot of time tracing the board before deciding it must be this way. I would be interested in knowing the reason for the preamp gain reduction though.
Now I'm going to try the other channel and give it a listen...
TDWesty said:
Hi Shawn,
It's PeAK chiming in. Good to see you on your way. The Bass controls attenuate the amount of feedback. With "no" feedback, the preamp gain would be the open loop value and be vary high. When 1/6 [560/(560+3300] the open loop output is fed back, feedback theory dictates that the gain is "roughly" the inverse of the "amount of feedback" or 6.
When R540 is removed, essentially 100% of the output signal is fed back and the gain attenuates to 1/1=1. Unity Gain. The 6 to 3 change I mentioned was due to a different mod related to a cap divider coming into the preamp section.
Not sure about the typo (have not verified) but it should be self evident as you have found. Last, when you are doing piecemeal mods, I'm not sure what you will get as the mods were broken down into sections but meant to done "together" so that issues with DC offset and improper biasing do not exist. If you do want to try things out, peice meal, then the feedback requires that feedback be done so in an AC sense and C524 be kept in at all times.
I'll try to check in here from time to time but feel free to email me and cross post here with your info for others to try.
P.S. I'll get around to modifying the broken links and uploading a component placement diagram...and fixing those typos you point out...
regards %^Q
cpemma said:
planet10 said:
I used cpemma link schematic thanks!
to set up NAD 3020i power amp in my Sim.
I wont tell you any results, of THD and Fourier ..
It will be my little
secret !!!!!So, what kind of power amplifier is NAD3020i ????
Really ???
As can be seen from my attachment
there is not any remarkable features.
Kind of 'old style', with bootstrap
and feedback to emitter of one single input BC550C.
So what makes this amp so loved, so good?
My opinion is:
- very well done compensation
- there are those small caps you know
- a wellbalanced topology, with
- separate V-Supply for Output and for Input VAS
- something I have recommended several times at forum
- and is the way I do it
see the input and voltage amplifier as a separate amp from output follower
Regards
lineup
Attachments
Hi,
that mexbro link to the NAD schematic is illuminating.
Nested feedback and lots of other designer tweaking shows in the complexity.
I wonder how long the designer spent optimising the component values!!!
For a cheap amp, the PSU is surprisingly complex and discrete opamps in the RIAA & Tone.
I begin to marvel at the value engineering that must have gone into all that.
No wonder it got good reviews, I can imagine how simple (=cheap to make and =cheap sound) it's competitors were in comparison.
that mexbro link to the NAD schematic is illuminating.
Nested feedback and lots of other designer tweaking shows in the complexity.
I wonder how long the designer spent optimising the component values!!!
For a cheap amp, the PSU is surprisingly complex and discrete opamps in the RIAA & Tone.
I begin to marvel at the value engineering that must have gone into all that.
No wonder it got good reviews, I can imagine how simple (=cheap to make and =cheap sound) it's competitors were in comparison.
...correction on comments
The 2 points above are either vague or wrong. I meant to say the components (R38 & R540) set the gain (and amount of feedback) above their turnover frequencies at mid to high frequencies. Then new feedback components (1k and 2.2k set the new gain at it is about 3).
The cap divider would actually attenuate some of the open loop high frequency gain. I've always found the NAD sort of muffled and the change allowed me to put in a better quality film unit with less HF roll-off when I was a poor student with no income.
PeAK said:
The 2 points above are either vague or wrong. I meant to say the components (R38 & R540) set the gain (and amount of feedback) above their turnover frequencies at mid to high frequencies. Then new feedback components (1k and 2.2k set the new gain at it is about 3).
The cap divider would actually attenuate some of the open loop high frequency gain. I've always found the NAD sort of muffled and the change allowed me to put in a better quality film unit with less HF roll-off when I was a poor student with no income.
planet10 said:
The amp came about in the days prior to CDP with their hotter outputs (1V versus 150mV) for sources. I bought into the better tracking with the muting on but eventually found that muting circuit introduce a phasey distortion. So despite years of alway using it, I found the sound improved without it even after I implemented the muting bypass. With the volume knob, the signal coming into the preamp is attenuated so the levels handle inside the high level section are reduced....possibly leading to less distortion.
Some notes...
Bak in the late Seventies(before NAD), Harmon Kardon was a big on current delivery and having separately regulated supplies for voltage gain sections. Solid state got a very bad wrap and HK were one of the first to hop on to Otala's research into transient distortion along with the NAD when it came out. When many were trying to outspec the other to impress bats, NAD believe in the school of "correct bandwidth limited" design...which means to only give a gain section a signal that it can handle.
The other aspect of the design is that they sought some of the brightest minds to consult. The phono section is a variant of Tomlinson Holman novel work with Advent on their phono section.
Wires and components in those days play a big part in the signature of the amplifiers (as they do today) but a major part of the sonic tuning was done by Bjorn Erik E...a lover and musician in a Jazz group at the time with a talent for circuit design by day.
As one reviewer put it, "The 3020 is a rose coloured frosted piece of glass"...but it provided the inspiration for the resurgence of Rotel and the founding of Creek.
lineup said:
Bak in the late Seventies(before NAD), Harmon Kardon was a big on current delivery and having separately regulated supplies for voltage gain sections. Solid state got a very bad wrap and HK were one of the first to hop on to Otala's research into transient distortion along with the NAD when it came out. When many were trying to outspec the other to impress bats, NAD believe in the school of "correct bandwidth limited" design...which means to only give a gain section a signal that it can handle.
The other aspect of the design is that they sought some of the brightest minds to consult. The phono section is a variant of Tomlinson Holman novel work with Advent on their phono section.
Wires and components in those days play a big part in the signature of the amplifiers (as they do today) but a major part of the sonic tuning was done by Bjorn Erik E...a lover and musician in a Jazz group at the time with a talent for circuit design by day.
As one reviewer put it, "The 3020 is a rose coloured frosted piece of glass"...but it provided the inspiration for the resurgence of Rotel and the founding of Creek.
Re: Some notes...
Thanks very much, PeAK
for telling us a bit of Amplifier history.
Interesting!
It is true the bandwidth of NAD 3020i is restricted.
Not only in local stages.
In my simulation it shows a total (including input lowpass filter)
upper limit -3dB at ~75.000.
But I can guess this was not too bad in those days when this amp was first released.
Still is not bad today.
It is surely enough for handling LF-signals within the audio band.
Regards
lineup
** LF = Low Frequency,
of which some frequencies are Audible to humans.
As opposite to HF, and VHF Very High Freq
PeAK said:
Thanks very much, PeAKfor telling us a bit of Amplifier history.
Interesting!
It is true the bandwidth of NAD 3020i is restricted.
Not only in local stages.
In my simulation it shows a total (including input lowpass filter)
upper limit -3dB at ~75.000.
But I can guess this was not too bad in those days when this amp was first released.
Still is not bad today.
It is surely enough for handling LF-signals within the audio band.
Regards
lineup
** LF = Low Frequency,
of which some frequencies are Audible to humans.
As opposite to HF, and VHF Very High Freq
I´ll jump in the thread with a question. Can anyone help me with my 3020 series 20. I can´t find the procedure for Idle Current Alignment since this version seems different from 3020 A-B-E-i and their service manuals.
This favorit of mine undergoes a minor upgrade with cans, main line resistors and cables. A thought to anyone of You clever fellows - The power transistors 2N3055/MJ2955 - Do they really wear with age and use? I have difficulty to understand need to replace them with new alikes. Is a new Toshiba for instance better than old original Motorola. Or is the meaning to replace them with modern equivalents?
Kind regards
This favorit of mine undergoes a minor upgrade with cans, main line resistors and cables. A thought to anyone of You clever fellows - The power transistors 2N3055/MJ2955 - Do they really wear with age and use? I have difficulty to understand need to replace them with new alikes. Is a new Toshiba for instance better than old original Motorola. Or is the meaning to replace them with modern equivalents?
Kind regards
Hi,
transistors appear not to wear out.
We break them by asking them to do too much.
Another thread just quoted an ONsemi MTBF value of 45million hours. That's just a tad longer than you or I will be here for.
I guess ONsemi used some statistical analysis to arrive at that, they certainly didn't measure it directly!
If the semiconductors are working properly then leave alone.
If you can design the amp to make use of alternatives then go ahead, but don't expect any sensible results to come from just swapping active components.
transistors appear not to wear out.
We break them by asking them to do too much.
Another thread just quoted an ONsemi MTBF value of 45million hours. That's just a tad longer than you or I will be here for.
I guess ONsemi used some statistical analysis to arrive at that, they certainly didn't measure it directly!
If the semiconductors are working properly then leave alone.
If you can design the amp to make use of alternatives then go ahead, but don't expect any sensible results to come from just swapping active components.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.