Acurus A250 and starground

[KNielsen]

@Doug Minderh wow, the noise issue aside, the aftermath is quite extreme... I really don't know after reading your findings where to address the problem so that your outputs will burn, besides from oscillation.
My first guess was a bad connection at the feedback (220uf), or just a tiny soldersplach that's almost invisible between the two solderpoints at this location that could bypass the cap or misdirected the signal??
Also it's a dual layer PCB, so maby one of the legs on the 220uf didn't make proper connection to the front layer copper if the cap moves just a bit when flipping the PCB ect ... This is possible, if when removing the old cap that the copper-insert in the PCB between the tracklayers leaves it's position together with the old cap... Check the old cap if this is a possibility, and if the insert had leaved together with the legs of the cap?? Personally I see duallayers where the designer choose frontlayer copper to a cap as a design flaw in terms of repairs afterwards. It can be tricky to remove enough solder to remove the cap without the insert follows. But often no problem......
At one point I also thought the possibility that you made an internal groundloop when applying signal and started probing, but normally the Zobel resistor will respond quite quickly if oscillation occurre and send magic smoke untill it burnes if one doesn't react fast and removed the source that courses the loop!
Lastly, did you confirm stable BIAS and mV over the emitters before testing, and that the pot is in good working condition? I don't think the pot is the real issue, but....

The headache is the burned powertransistors. This I haven't seen in "testmode" unless someone forgot to use white thermal paste and insulation between the heatsink and tanies....

Best regards

 

[Doug Minderh]

@Doug Minderh wow, the noise issue aside, the aftermath is quite extreme... I really don't know after reading your findings where to address the problem so that your outputs will burn, besides from oscillation.
My first guess was a bad connection at the feedback (220uf), or just a tiny soldersplach that's almost invisible between the two solderpoints at this location that could bypass the cap or misdirected the signal??
Also it's a dual layer PCB, so maby one of the legs on the 220uf didn't make proper connection to the front layer copper if the cap moves just a bit when flipping the PCB ect ... This is possible, if when removing the old cap that the copper-insert in the PCB between the tracklayers leaves it's position together with the old cap... Check the old cap if this is a possibility, and if the insert had leaved together with the legs of the cap?? Personally I see duallayers where the designer choose frontlayer copper to a cap as a design flaw in terms of repairs afterwards. It can be tricky to remove enough solder to remove the cap without the insert follows. But often no problem......
At one point I also thought the possibility that you made an internal groundloop when applying signal and started probing, but normally the Zobel resistor will respond quite quickly if oscillation occurre and send magic smoke untill it burnes if one doesn't react fast and removed the source that courses the loop!
Lastly, did you confirm stable BIAS and mV over the emitters before testing, and that the pot is in good working condition? I don't think the pot is the real issue, but....

The headache is the burned powertransistors. This I haven't seen in "testmode" unless someone forgot to use white thermal paste and insulation between the heatsink and tanies....

Best regards

Thanks for helping. I used a "solder sucker" to remove solder before pulling the components. I didn't have to force any of them out, perhaps just hold a little heat. I have one of the very large solder suckers that actually works.

I also tested the pads I worked on for continuity before powering on.

I did not confirm bias or the bias pot.

Can you confirm that the 220mfd cap is indeed bi-polar? I'm wondering if that allowed the circuit to go into oscillation for some reason...

My theoretically new old stock transistors will be in at the end of the week. I feel like I need to trace the circuit out and make a schematic, then thoroughly test all of the components before putting any of the power transistors back on the board.

 

[KNielsen]

@Doug Minderh This should be the schematic your searching. It states 220uf bi-polar just like you installed.
It's the DIA100-model, so the input section may be different, but the A100-schematic is in poor quality.
I would start with a meter over each component to confirm their state before further repair

Regards

 

[Jeandelyon]

J'ai joint une image d'un des tests précédents que j'ai effectués sur un A250 sur lequel j'ai travaillé. Ces données ont été collectées et saisies manuellement dans une feuille de calcul avec un ancien ensemble de test analogique que j'ai. J'utilise maintenant un système Audio Precision, ce qui me permet d'obtenir de bien meilleures informations. Ce test a été effectué avec un filtre LP de 30 kHz qui maintient la distorsion un peu au-dessus de 10 kHz. Pour obtenir un test comparable, vous devez donc faire de même sur le système de test que vous utilisez. Si vous utilisez 80kHz BW ou 20kHz BW (ce que vous semblez avoir utilisé), les tests seront faussés par rapport à celui que j'affiche. Je testerai un autre A250 sur l'AP S1 dans le mois prochain et je pourrai publier cette information lorsque j'aurai terminé.

I don't think your numbers are that far out of line with what I've measured in the past. My past tests are a bit higher, but I'm on 60Hz power which will play a role. 48Vrms = 135Vpp and once you factor in power supply drop when the amp is that heavily loaded your right in the region that you should be. Remember that 90.2Vdc is with almost no load on the amplifier. The next time you test the amp clip your DMM onto one of the power supply capacitors and measure the DCV when you do a full power test to see how much the power supply rails are dropping when under test. You also have to factor in how much the main AC line is dropping when the amplifier is drawing this much power from the wall and its certainly dropping several volts when fully loaded.

View attachment 1418516

hello,
I think I'm talking to the right person, given the technicality of your remarks! I have an Acurus 125 x 5 which uses a lot of elements from the 5x200. My issue is the following . My amplifier is starting to generate a weak crackling sound (mids, treble) in all 5 channels. This is not a "hum" or "buzz" type sound, but more like non-periodic crackles (not 50 or 100 Hz). Do you know this problem, and, what would be the components to check on the power supply part I suppose? Thank you for your helpful help ;-)

Looking good there. You decided to go all out redesigning the rear panel is a lot of added work for sure, but It'll look nice. You should drill the capacitor mounting bracket holes before sending the lower chassis out for painting.

Bonjour, je pense m'adresser à la bonne personne, vu la technicité de vos propos ! J'ai un Acurus 125 x 5 qui utilise beaucoup d'éléments du 5x200. Mon problème est le suivant. Mon ampli commence à générer un faible grésillement (médiums, aigus) sur les 5 canaux. Ce n'est pas un son de type "hum" ou "buzz", mais plutôt des craquements non périodiques (pas 50 ou 100 Hz). Connaissez-vous ce problème, et, quels seraient les composants à vérifier sur la partie alimentation je suppose ? Merci pour votre aide précieuse ;-)

 

[mondialfan]

hello,
I think I'm talking to the right person, given the technicality of your remarks! I have an Acurus 125 x 5 which uses a lot of elements from the 5x200. My issue is the following . My amplifier is starting to generate a weak crackling sound (mids, treble) in all 5 channels. This is not a "hum" or "buzz" type sound, but more like non-periodic crackles (not 50 or 100 Hz). Do you know this problem, and, what would be the components to check on the power supply part I suppose? Thank you for your helpful help ;-)

The 125X5 uses 5 discrete output boards, one for each channel and a single shared power supply. So if the noise is in all 5 channels then your issues are not related to the channel boards due to the fact all 5 of them haven't failed at the same time. The power supply in this amp is a basic x-frmr/bridge rectifier/pair of e-caps with no regulated stages. This also does not sound like a power supply problem that I have ever experienced. I would look more closely at the preamp/processor or source components as a possible source of the issues.

 

[Jeandelyon]

The 125X5 uses 5 discrete output boards, one for each channel and a single shared power supply. So if the noise is in all 5 channels then your issues are not related to the channel boards due to the fact all 5 of them haven't failed at the same time. The power supply in this amp is a basic x-frmr/bridge rectifier/pair of e-caps with no regulated stages. This also does not sound like a power supply problem that I have ever experienced. I would look more closely at the preamp/processor or source components as a possible source of the issues.

Thank you very much for your reply !
If I use an other 5 Channels amplifier at the place of the ACURUS , I have no sound issue .

 

[KNielsen]

Long time since my last post here... I've almost stopped building due to lack of parts, or the long wait for parts! Nevertheless after seeing a new L10-thread here I went to my office too see what I left behind form the past..... So here goes:

I was advised to skip D1/2 and D4/5 in my A250 at the current source in favour of a LED diode (The LED's should sound better some posted in another thread a while ago.. I cannot see why at this location, but it's worth a shot with some spare time at hand, so I will not neglect this option before I've tried?!)

To get some ~3.6mA across the LED's I need to lower R4 from 200Kohm to around ~45Kohm or lower, but this will also increase the flow across R4 to aprox 4mA (simulation): is this a problem?
The "flow" across the original 200Kohm is a couple of 100's μA

R5 and R12 should be adjusted accordingly... I've made some simulations as written, but have to measure the constant current source in my A250 just to be sure before I'll continue, so I get the right amount mA from the source.



Best regards

 

[Fsatsil]

Hey KNielsen sorry I can’t offer any help with your Led current source project but it’s nice to see that I’m not the only one still messing around with these old Acurus amps! Id really like to add speaker protection to my 2ch 200x3 so I have a little peace of mind when I have the Acurus connected to expensive speakers. Haven’t decided if I want to spend the money on one of the fancy solid state relay protection boards or just a basic standard relay type. Anyone have any advice?

 

[mondialfan]

Long time since my last post here... I've almost stopped building due to lack of parts, or the long wait for parts! Nevertheless after seeing a new L10-thread here I went to my office too see what I left behind form the past..... So here goes:

I was advised to skip D1/2 and D4/5 in my A250 at the current source in favour of a LED diode (The LED's should sound better some posted in another thread a while ago.. I cannot see why at this location, but it's worth a shot with some spare time at hand, so I will not neglect this option before I've tried?!)

To get some ~3.6mA across the LED's I need to lower R4 from 200Kohm to around ~45Kohm or lower, but this will also increase the flow across R4 to aprox 4mA (simulation): is this a problem?
The "flow" across the original 200Kohm is a couple of 100's μA

R5 and R12 should be adjusted accordingly... I've made some simulations as written, but have to measure the constant current source in my A250 just to be sure before I'll continue, so I get the right amount mA from the source.

View attachment 2613710

Best regards

You'll need to use a big resistor for the 45k, I'd use a 2W rated resistor. You'll be dropping on the order of 2V across the LED versus the 1.3 - 1.4V for the existing 4148 diode pair so yes you'll have to increase the R at R5/R12 to maintain the original current. I really do not see a big difference being made here to sound quality by installing the LED in place of the 4148's.

 

[KNielsen]

Hey KNielsen sorry I can’t offer any help with your Led current source project but it’s nice to see that I’m not the only one still messing around with these old Acurus amps! Id really like to add speaker protection to my 2ch 200x3 so I have a little peace of mind when I have the Acurus connected to expensive speakers. Haven’t decided if I want to spend the money on one of the fancy solid state relay protection boards or just a basic standard relay type. Anyone have any advice?

To my knowledge, all the Aragon's are the relaytype, and works very well.

I've drawn the original schematic from the Aragon 4004mk.l, as seen here if you wanna build a speaker protect yourself that somewhat follows the first Mondial amps.
I haven't been able to buy the original relay, so keep an eye on Mouser or Digikey for stock

R100 is connected to the secondary at the xformer, and should be raised in value to meet the higher voltage in the A250 over the 4004 PSU



Best regards

 

[Fsatsil]

Thanks!

 

[KNielsen]

Thursday evening... Some benchtesting before the amp is taken apart and restored.
-No upgrades (mods) on this one, only recap and new resistors at input, R7/14 and feedback and zobel, and BIAS pot.

It's been a long time since the last Acurus amp on my table, but I really do enjoy when it happens. Parts from Mouser also arrived today together with the A250, so it's just getting things started.

I did however find that I had some "scope" problems when measuring both left and right channel at the speaker output +8omh loads on my 2-channel scope. Seamed like a ground issue on the scope interfering, so I ended up testing only one channel at the time instead of multiple when sweeping through the sine and squarewaves..
Might have to dig into this a bit further for future probings, and learn (read) a bit more on subject... My probes are the stock ones that came with the Siglent o.scope, and I'm unsure if a resistor In-between could be the solution

Best regards

 

[KNielsen]

Done:
1kHz/5Vpp square wave is spot on
1kHz/10kHz/100kHz sine wave are perfect.
I didn't get more into the conflict between channels on my scope as I'm afraid this might pull the Zobel, but it did 320watts at 8ohms/1kHz one channel driven. I'll suspect to see 2x290 on both if I dare..
Voltage at the rails is 85VDC precise.
DC-OFFSET is the best I've seen so far: 0.2mV (L.ch) and 3mV (R.ch).
I have really high expectations for the quality of the sound when music is applied tomorrow

A little about the amp:
It's one of the later models, black Toshiba throughout, the signal GND is like GEN2 directly to PSU and not Speaker minus.
Still there were some part mismatch between the two circuits, black caps in one Vs Wima in the other.
It's all WIMA MKP4/10 and MKS10 now, probed to set the the outer foil to GND. I have not done this before hens the small values of the WIMA's, so I'm quite curious if the sound reveals my effort

Best regards

 

[Andersonix]

There's so much empty space in there that it would be easy to make banks of smaller caps (2x4?) to connect to the posts of the existing two caps and really get the PS rail ripple down, so you could hear that silence. If you were ambitious you could even decouple the PS rail fuses on each pcb.
(And once the PS is quiet enough, you can dial in some bias!)

 

[KNielsen]

There's so much empty space in there that it would be easy to make banks of smaller caps (2x4?) to connect to the posts of the existing two caps and really get the PS rail ripple down, so you could hear that silence. If you were ambitious you could even decouple the PS rail fuses on each pcb.
(And once the PS is quiet enough, you can dial in some bias!)

This is a far as an A250 could be build to my knowledge.. It requires some serious cutting in the chassis and cannot be undone once performed. 8x13.000uf in a pseudo dual-mono setup.
I would never skip the fuses though

Best regards

 

[Andersonix]

Maybe 1000uF on the main pcb is enough to decouple the fuses?

But I'm not suggesting you cut into yours, but leave it totally reversible for now: Just hanging a cap bank (or two) via the ring connectors, so you can hear it and evalutate it... and then decide!

 

[KNielsen]

Maybe 1000uF on the main pcb is enough to decouple the fuses?

But I'm not suggesting you cut into yours, but leave it totally reversible for now: Just hanging a cap bank (or two) via the ring connectors, so you can hear it and evalutate it... and then decide!

Maby I do not understand, but I build and own the last amp posted with big help from @mondialfan in this thread

 

[mondialfan]

Maybe 1000uF on the main pcb is enough to decouple the fuses?

But I'm not suggesting you cut into yours, but leave it totally reversible for now: Just hanging a cap bank (or two) via the ring connectors, so you can hear it and evalutate it... and then decide!

We have implemented the rail decoupling caps after the fuses, I termed it "post fuse DC rail caps". You can read about as early as page 1 in this thread.

 

[Andersonix]

Yes, now I see it Mondialfan! 100uF right?

 

[mondialfan]

Yes, now I see it Mondialfan! 100uF right?

Yes those are the ones.