Sansui AU Alpha 707MR total recap: Elna Silmic II or Nichicon

Do you want to replace the main filtration ???
then replace the amplifier ...
do not move it !!!! They are okay !!! This eggy Nichicon will survive you, as does Silmic who is with him ...

With all due respect - there is nothing worse than to mention the main filtration or transformer in Sansui ...

Or let's do it - how do you replace it, those removed from 707 send me OK ????
 
Hi all, I'm new to this site, although I have browsed threads in the past.

I'm a hobbyist with guitar amps and pedals and audio gear. I wanted to jump in on this if I may because I have a Sansei AU Alpha 607 MRX which I love. After lots of other amps including Naim, AudioNote and Unison Research gear, I find this amplifier so natural and enjoyable listen to. All that said I just recently got tempted to tinker; Thinking the bass was probably softer than it should be I changed /uprated the main reservoir caps to 15000uf nichicon super through (the originals being type 2 12000uf caps). I had not read @czubsi 's comments about setting the amp up first and not changing the reservoir caps unfortunately. I agree that maybe the biasing would have fixed this, but whilst at first it sounded like crap, after a day or so of run in its starting to sound significantly tighter, bigger in the bass, and the upper end its starting to smooth.

My amp came with mainly unicon electrolytic caps and a few chemicon. Because of this I also changed all the supply board electrolytics to Elna Silmic ll. I did this because other pictures of the AU alpha 607 MR/MRX had Elna Silmic and the spacing on the boards fits these caps better than the unicon in some places. I'm waiting to see how these sound after running in before I change the signal path electrolytic from unicons to Elna.

What I loved about the amp was its naturalness especially in high end and midrange detail. I wonder if some of this was due to it being biased cooler due to drift in values, although the union/chemicon power board caps (which I've kept) were all pretty much spot on.

The amp is coming round and starting to sound nice again with much more bass weight, but my question @czubsi, if he might be so kind, would be, how do I set this up? Had I read your comments before I would have asked this first. I have built and biassed plenty of valve amps. I see there is a bias trim pot on each power board along with two hot/cold trim pots. Where would i meter to read and adjust these and where would i find the correct readings? Would these be milli volt or milli amp measurements?

Any help from @czubsi or anyone else who knows would be very much appreciated.

Thanks in advance,

Carson
 
Here's a link to the 607MR/MRX service manual which has the info you need:

https://www.hifiengine.com/manual_library/sansui/au-a907mr.shtml

Get the Elna Silmic II caps in there pronto! I did a previous Silmic II upgrade on an AU-X711 (same general topology as the 607MRX) and she sounds lovely. I will do the 607MRX in the near future.

Thanks Pete,

That is extremely helpful! I am cautious because after years of trying different gear, I love this amp. That said after two days burn in. I'm fairly confident that you are right and the silmic ll in the signal path will be a further improvement. I'm in Italy visiting my brother at present so will attack the power amp boards upon my return.

I did have a few questions regarding the set up instructions though:-

Have you set up your 607 MRX using the manual? After having a read through, I noticed that 3-2-2 step 2 says to connect the probe between cold(speaker) terminal and cold (speaker) terminal. I'm assuming this must be an error as there would obviously be no potential difference between the same terminal.
Looking at 3-1-2 step 2 for the 907, it seems it should say between cold and hot (speaker) terminal?

Also, am I right to assume that the use of '(speaker)' after hot and cold in some but not all cases, does not mean metering to the actual speaker output terminals, but rather, just means the hot and cold out terminals on the board?

3-2-2 step 1 instructs metering between hot out and GND, where as 3-1-2srep 1 (for the 907) instructs metering between hot out and hot bias, for the same function, as there appears to not be a ground test point on the 907 board; Is this difference correct.

The final question is just regarding the recomnded 'pre heating'; Is this necessary, as I don't have the required test bench gear to generate a 20 kHz sine wave or stereo dummy load resistors? I do have mono attenuators and dummy load for guitar amps and could and may have close to 20KHz test signal on a set up CD that I could loop via my CD player. I'm also unsure if, in this case, the metering of the volume level in volts is from the actual speaker terminals or the hot and cold out on the board? Could I just run the amp for 30 minutes to warm it up and then carry out the adjustments with my usual speakers connected (nominal 8 ohm load)?

Any advice would be appreciated, although I will just carry out a rough check based upon my suspicions otherwise as it sounds like it is biased right.

Thanks once again,

Carson.
 
Also, am I right to assume that the use of '(speaker)' after hot and cold in some but not all cases, does not mean metering to the actual speaker output terminals, but rather, just means the hot and cold out terminals on the board?

3-2-2 step 1 instructs metering between hot out and GND, where as 3-1-2srep 1 (for the 907) instructs metering between hot out and hot bias, for the same function, as there appears to not be a ground test point on the 907 board; Is this difference correct.
There are many errors in the service manuals (just like you spotted measuring between hot and hot bias)
adjustments u have to do:
hot bias on special test pins
cold bias on special test pins
hot speaker terminal to cold speaker terminal with selector set to balanced and vol to minimum
hot speaker terminal to ground (actually dont spend too much time on this adjustment as it can'not be adjusted perfectly)
hot speaker terminal to cold speaker terminal with selector set to integrated and vol to maximum, in this case the dc offset of the preamplifier will be added to the power stage, trim pots for this adjustments are behind front panel.
 
I realize this, this adjustment keeps wobbling... For what reason this adjustment does not stabilize?
Because the stability depends on the overall stability of the secondary power supply feeding the power to the voltage amplifier, as you may know the current stage is supplied with different supply than volvate amplifier.
This adjustments affect reference voltage of the constant current transistors of the negative rail of power supply, this voltage is provided fot both channels, keep in mind that you are adjust the difference of mV.
Actually it doesn't affect anythin related to the sound reproduction, you cant notice the difference when positive power supply rail will be +38.15V and negative will be -38.01V
 
Hi,

Just read this thread properly for the first time. I don't want to chime in too much on the issue of flavour of different caps except to say most differences will be in the listeners mind (this is not too say caps don't sound different but the differences should be subtle unless a part is faulty, with the exception of signal coupling caps).

What I do want to chime in on is that I have inspected the main filter caps of an AU-Alpha607MR and compared them to the best caps I could find to fit. None of the new replacement parts are in the same league as the original in terms of ESR and phase angle. The original is an order of magnitude better in both of these than the best modern PS caps I could find. So I agree with the previous statements to be hesitant with replacing those. This is coming from a guy who rips out power supply caps from (solid state) amps as standard practice.
 
Last edited:
Exactly like that, i have wrote the same on a polish audio forum, realize that main filtering caps in 907MR have ESR below measurment limit of my RCL Bridge...

secondary, everything in sansui alpha amplifiers work on theuir 20-30% abilities, just check the supply voltage and used capacitors, power supply of 36V and capacitors rated of 85V, no worries about capacitors for the next 30 years.

btw, even in lower grade amplifiers, i have serviced maybe 1000 of them and i have replaced maybe 100 capacitors in my whole carrer. when i got amplifier in service the first thing and the most important is to clean every switch and potentiometer as there is the most signal drop out.
 
Thanks!

BTW, would you know what secret behind the choice of PS reservoir caps by Sansui?
For example, the alpha 707MR power supply has a 5600uF and a 8200uF main caps (total 13800uF only).
That story that "the more, the better", does not seem to make sense to the Sansui designer. The lower model, alpha 607MR has 2 caps of 12000uF (total 24000uF).
 
I will also add that after replacing the AU-Alpha607MR power supply caps with replacements - the best of the bunch I bought at great expense - the power supply ripple when under stress (demanding output) was about twice twice as bad as with original caps.
 
Thanks!

BTW, would you know what secret behind the choice of PS reservoir caps by Sansui?
For example, the alpha 707MR power supply has a 5600uF and a 8200uF main caps (total 13800uF only).
That story that "the more, the better", does not seem to make sense to the Sansui designer. The lower model, alpha 607MR has 2 caps of 12000uF (total 24000uF).

in the end, the total power output available is limited by the total capacitance (and transformer etc) but the value used on the 707 is already enough for 130W. However, transient response of the power supply reservoir caps is more complicated and as a result this is why you see a cascade of cap values in the power supplies, which results in better ability of power supply to handle transients (the technical details are long winded and depend not only on Capacitance value but other factors).

As for why the 707MR appears to have less C than the 607MR I don't know the answer except they must be relying on transformer. The 707MR has one oval cap, which Sansui claimed was a bit special (and can be found in the high end alpha models). I am guessing the large caps on the 607MR are partly a result of putting what will fit without having to make a new chassis part. Maybe someone else knows more about this quirk.
 
Thanks!

BTW, would you know what secret behind the choice of PS reservoir caps by Sansui?
For example, the alpha 707MR power supply has a 5600uF and a 8200uF main caps (total 13800uF only).
That story that "the more, the better", does not seem to make sense to the Sansui designer. The lower model, alpha 607MR has 2 caps of 12000uF (total 24000uF).

and 907MR has 5600+3900 per side (total 9500uF)

the higher amp is the smaller capacitance it has, but its not only about C, but mainly about ESR which depends of the ability to feed the power transistors on demand.

everything depends on the transformer power, you need to know that in theory you need 1000uF for each 1A of current for efficient filtration.
but capacitor except filtration has ability to magazine energy, and as you may know capacitors are much more cheaper than power transformers.

thats why TOTL amplifiers from sansui have huge power transformers of total power about 1.3kVA and only about 9500uF per side.

whats it makes to the sound? from my experience sound coming out from amplifier with small capacitors and big trafo is faster, ethereal and much more distributive.
 
and 907MR has 5600+3900 per side (total 9500uF)

the higher amp is the smaller capacitance it has.

everything depends on the transformer power, you need to know that in theory you need 1000uF for each 1A of current for efficient filtration.
but capacitor except filtration has ability to magazine energy, and as you may know capacitors are much more cheaper than power transformers.

thats why TOTL amplifiers from sansui have huge power transformers of total power about 1.3kVA and only about 9500uF per side.

whats it makes to the sound? from my experience sound coming out from amplifier with small capacitors and big trafo is faster, ethereal and much more distributive.

Great info. On the last sentence, I would propose that big transformer and many small caps is even better than big transformer with just a few small caps. my favourite big amp (an not very well known, outside of Australia, ME-1400 power amp) had a 5kVA transformer and over 90x 3000uF power supply caps. I spoke to the designer and he talked about various impedance advantages as well as smaller caps handling transients better.
 
Last edited:
why are you guys busting out our nuts with facts and information? I was much happier thinking that gigantic PS caps meant super fidelity hi-fi sound. :(
So the $100 I spent on 4x10000uF 80V Nichicon Gold Tunes caps in my AU-D11 Mk2 refresh was not needed?
 
Like i said, 1000 amplifiers, maybe 100.000 of measured capacitors and i just changed few of them. Usually 20-30 years of work doesnt affect their ESR and capacitance.
While servicing an amplifier there are better places to pay attention than capacitors.

1. Cold joints
2. Cleaning switches and potentiometers
3. Change thermal paste
4. Adjustments
 
and 907MR has 5600+3900 per side (total 9500uF)

the higher amp is the smaller capacitance it has, but its not only about C, but mainly about ESR which depends of the ability to feed the power transistors on demand.

everything depends on the transformer power, you need to know that in theory you need 1000uF for each 1A of current for efficient filtration.
but capacitor except filtration has ability to magazine energy, and as you may know capacitors are much more cheaper than power transformers.

thats why TOTL amplifiers from sansui have huge power transformers of total power about 1.3kVA and only about 9500uF per side.

whats it makes to the sound? from my experience sound coming out from amplifier with small capacitors and big trafo is faster, ethereal and much more distributive.
Very interesting. I always see amplifiers being "judged" by the amount of capacitance: the more, the better... And I've been intrigued for a long time with the Sansui alpha, which, in addition to having modest capacitance value, also combined capacitors of different brands and values, in parallel, which is not very common to see. Well, it seems that all this peculiarity makes sense, and Sansui knew very well what it was doing. The sound of a Sansui alpha is something very special.
 
I will explain like that, capacitors are like turbo in car
turbo givex you boost just like capacitors but the travel from capacitor to power transistor is delayed due to ESR, and in engine with turbo you have poor reaction for gas

Big transformer = litres in engine, its from where come the power.

also, think about electric circiuit in your home, all the power comes from bigga transformer somewhere in your town (ofc capacitors can'not be placed here)
 
Back
Top Bottom