C33 Update Project 2013

[Bubo]

That might be good in the lab.. but this isn't `beater gear´ :bash: .. man!

LOL - ok we're laughing here!



-Gregory

Done with care, a little capacitor rolling shouldn't hurt anything if they are all withing spec. I'm pretty sure the design engineers experimented with components before making their final selections and didn't rely solely on spec sheets and board level testing. Or at least I hope they didn't.

I know that in the telecom gear we brought in every viable component option for new tech and tested every one of them for function and stress. The results weren't predictable.

The one that interested me the most were the dsps doing voice compression and decompression and the competing algorithms, sound quality under ideal and non ideal was always interesting to listen too, the acid test. The was done in 1 to N packet node networks with multiple routs, in the lab with introduced delay to simulate transoceanic and multi-satellite hops.

The packet switches could do it, but the TDM Telco backbone would not have been able to keep up with the speeds and tech they had at the time.

 

[jlovda]

Done with care, a little capacitor rolling shouldn't hurt anything if they are all withing spec. I'm pretty sure the design engineers experimented with components before making their final selections and didn't rely solely on spec sheets and board level testing. Or at least I hope they didn't.

I know that in the telecom gear we brought in every viable component option for new tech and tested every one of them for function and stress. The results weren't predictable.

The one that interested me the most were the dsps doing voice compression and decompression and the competing algorithms, sound quality under ideal and non ideal was always interesting to listen too, the acid test. The was done in 1 to N packet node networks with multiple routs, in the lab with introduced delay to simulate transoceanic and multi-satellite hops.

The packet switches could do it, but the TDM Telco backbone would not have been able to keep up with the speeds and tech they had at the time.

I think you're dealing with comparisons of apples and oranges here. The world of DSPs is one of algorithms and IC speeds to determine frequency response and distortion. If you examine vintage Mcintosh equipment, the physical circuitry basically looks all the same. Certain types of components consistently show up in the same places. The same Mallory, metal can twist prong electrolytics tend to be in the power supplies, etc. One model may have mostly Nichicon in it while another earlier or later model may have Rubicon. I asked a question to Ron-c a year or so ago as to who Mcintosh didn't use expensive Nichicon Muse or film capacitors in their gear if they were a high end manufacturer. He stated they use high quality "audio grade" caps that meet the required specs of the unit. Back then they did test transistors for gain and noise. We are NOT talking about today's IC laden room equilization devices. A company like Mcintosh is not in a position to have three IC design houses come up with samples.

 

[Bubo]

You probably could eliminate the input caps if you were sure there was no DC coming in from the source but don't trust me without checking somewhere else as well.

Assuming zero DC, the caps are there strictly as a low pass filter? If more than the 20khz gets past them, is there an issue or was this a design goal number that someone plugged in?

We are NOT talking about today's IC laden room equilization devices. A company like McIntosh is not in a position to have three IC design houses come up with samples.

If it makes my argument more palatable, substitute "any component" for DSP. I'm pretty sure more than one type of cap was brought in for the selection process and that a systems test that included the subjective how does it sound was part of the process. I would be stunned if it wasn't.

Another factor is component cost management, sometimes needlessly so.

 

[jlovda]

Assuming zero DC, the caps are there strictly as a low pass filter? If more than the 20khz gets past them, is there an issue or was this a design goal number that someone plugged in?

Mcintosh cannot guarantee there will not be DC at the input terminals.

Regarding selection of parts, Ron-c might have the definitive answer. One thing I will say at the risk of being kicked off this board or shunned by Ron is that Ron has NEVER (as far as I remember) used subjective terminology such as "soundstage," "imaging," etc in responses. The specs you see at the back of the owner's manual are are the targets the company is shooting for.

 

[Bubo]

Mcintosh cannot guarantee there will not be DC at the input terminals.

Regarding selection of parts, Ron-c might have the definitive answer. One thing I will say at the risk of being kicked off this board or shunned by Ron is that Ron has NEVER (as far as I remember) used subjective terminology such as "soundstage," "imaging," etc in responses. The specs you see at the back of the owner's manual are are the targets the company is shooting for.

Design goal specs are simply a numerical representation of what people think they want, whether they actually want it that way is another matter.

"We built what you asked for, not what you needed"

Short version, a design goal that I gave a mechanical engineer was that a telecom system cabinet 4ru had to be able to survive a drop off of a loading dock with no cards but the backplane installed.

My requirements for that cabinet were something like 4 or 5 sentences.

He wrote his own specs and had a lot of fun doing it, I think the shipping guys had some fun too. The customers got what they wanted, gorilla proof cabinet and box.

My requirements for a stereo system would probably read something accurately reproduce a specif set of known high quality sound recordings using certain sources, and certain receiving units external to the system, or at least allow tweaking for the end of the chain. For our equipment one goal would have been "glass front bezels must be customer installable in less than 5 minutes and must be able to survive a drop in their own packaging", and the secretaries would have been the mock customers.

There is always the spec out factor in any sales process, but an informed sales force can usually get past the nonsense easily.

My 2 cents.....

 

[dubai2000]

Well, in the spirit of discussion I do believe that different caps might influence the sound of any preamp/amp etc - depending on where they are in the circuit.
Personally I do have certain brands that I don't use as the 'tone' they produce is not pleasant to my ears - knowing quite well thate here we enter the realm of subjectivity. A 2.2uF cap from brand x should measure like a 2.2uF cap from brand Y (looking ony at capacity at this point). Still, my ears tell me that they might sound different - and that is for me the reason why I might prefer a cerain type/brand over another one.

Personally I have chosen Elna Silmic II caps for the signal path and Panasonic FM (not FC!) for the power supply part.

In the upcoming Christmas vacation I might try removing those caps on the input board - it just seems to be a PITA to remove those.......

Wolfram

 

[Gregory]

Well, in the spirit of discussion I do believe that different caps might influence the sound of any preamp/amp etc - depending on where they are in the circuit. In the upcoming Christmas vacation I might try removing those caps on the input board - it just seems to be a PITA to remove those.

Quick report on the sound of recaps

Been playing the C33 over the past 2 or 3 days through its own Monitor Amp and like what I'm hearing very much! Starting in the afternoon around 3 pm and running till roughly midnight when our 'video' ends. Showtime is typically 9:30 pm.

I'm hearing pleasant detail in both acoustic and electric strings, and the streaming Christmas music is delightful. It fills the house.. up the spacious folded staircase to the upstairs office/library and bedrooms with serene instruments.

Also using a good measure of the 4th and 5th C33 EQ controls to `roll off´ some of the noisy rock and pop music of past decades, but for clean holiday instrumentals it's flat bandwidth all the way up. Also the 30Hz C33 tone control can function as the CS3.5 bass EQ up to ~90° or 3:00 position before the bandpass curves diverge too much.


Your McIntosh Integrated

You plan to REMOVE the caps from your MA 6200? You mean take them out of circuit and wire direct? It's not that hard to remove the input boards on the C33 for example. Just unplug the connector and remove the screws from the back of chassis. Best to take digital photos on make sketches for reference as we did in the 'old' days. Be in the right mood and take it slow. It should be enjoyable. If you're PO'd then wait for another day.


-Greg

 

[Bubo]

Quick report on the sound of recaps

Been playing the C33 over the past 2 or 3 days through its own Monitor Amp and like what I'm hearing very much! Starting in the afternoon around 3 pm and running till roughly midnight when our 'video' ends. Showtime is typically 9:30 pm.

I'm hearing pleasant detail in both acoustic and electric strings, and the streaming Christmas music is delightful. It fills the house.. up the spacious folded staircase to the upstairs office/library and bedrooms with serene instruments.

Also using a good measure of the 4th and 5th C33 EQ controls to `roll off´ some of the noisy rock and pop music of past decades, but for clean holiday instrumentals it's flat bandwidth all the way up. Also the 30Hz C33 tone control can function as the CS3.5 bass EQ up to ~90° or 3:00 position before the bandpass curves diverge too much.


Your McIntosh Integrated

You plan to REMOVE the caps from your MA 6200? You mean take them out of circuit and wire direct? It's not that hard to remove the input boards on the C33 for example. Just unplug the connector and remove the screws from the back of chassis. Best to take digital photos on make sketches for reference as we did in the 'old' days. Be in the right mood and take it slow. It should be enjoyable. If you're PO'd then wait for another day. :thmbsp:



-Gregory

Good news, in addition to the re-cap have you considered updating the power supply to increase both efficiency and tighten the tolerances of the output? I haven't looked, but I am guessing there are a bridge and zener diodes the could be "improved"? Is the transformer isolated from the chassis in this unit?

I would probably put off the platinum power cord woven by singing virgins under the moon light of the extra day on a leap year, and keep that in the pocket for the last 1%:scratch2:

Once stabilized, how are you going to measure the before and after on the Op amp rolling? I haven't tried it, but the headphone guys swear by it.

 

[Bubo]

Well, in the spirit of discussion I do believe that different caps might influence the sound of any preamp/amp etc - depending on where they are in the circuit.
Personally I do have certain brands that I don't use as the 'tone' they produce is not pleasant to my ears - knowing quite well thate here we enter the realm of subjectivity. A 2.2uF cap from brand x should measure like a 2.2uF cap from brand Y (looking ony at capacity at this point). Still, my ears tell me that they might sound different - and that is for me the reason why I might prefer a cerain type/brand over another one.

Personally I have chosen Elna Silmic II caps for the signal path and Panasonic FM (not FC!) for the power supply part.

In the upcoming Christmas vacation I might try removing those caps on the input board - it just seems to be a PITA to remove those.......

Wolfram

I did read a write up by a group of engineers, perhaps in France, that did a capacitor shoot out based solely on their subjective listening experience and I want to say they would agree with your choice. It has been a while since I read it.

I did notice they were holding wine glasses, which in the past has improved my listening experience; but it wasn't with a bunch of other male engineers in the room.:yes:

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Have you thought about how to characterize the output for a before and after report?

Walking a test tone up and down the frequencies is good for one frequency at a time, but under a complex audio signal, how would you say my "radio receiver is tuning every station at once perfectly" (best analogy I can think of on 2 cps of coffee)?

Program the receiver(s)so it knows where to look for it's signal?

I suppose you could use multiple signal generators (string quartet emulation), combine the output, and measure the before and after as they cycle up and down their frequency ranges? It would require some processing power to keep track of the signals, or at least attempt to.

One test I would want to see is the behavior of the pre-amp when 4 string instruments are all brought into a narrow intense range.

I know how to do this in the data world, but not sure in pure analog with sine waves.

I am not familiar with the test equipment, but suppose you could loop their outputs through the test subject, and program it to know were to look for it's signal (as it cycles like a string instrument) and keep a log of the variations. Start with one test signal, then two and so on. :scratch2:

 

[Gregory]

Compandor section recap

Had enough caps together to start on the left-rear section of the main board.

This is a BEFORE view.. and check out those High Level Input caps on top!


Original Compandor caps pointed out on main board, and review of High Level Input caps. Image originally posted 12-19-2013.


New caps installed

New Compandor caps on main board are a mix of Nich ES, FG, KZ and VR, VZ.


Updated caps highlighted above.


Compandor SPEED caps

These three caps were replaced with black Nich VX and VR caps that were on-hand.

KZs were used for C509-510 on the Input stage of the VCA's because they were bought in 2009 and needed to be used-up!


Are those two BIAS pots on the left dissimilar?




C509-510 (KZ replacements above) desoldered and removed..




A closer look at the COMPANDOR SPEED caps.

I think the largest of them was in backward. The new cap was put in according to the schematic.




Input coupling caps and VCA drive

Two 1uF non-polarized VCA Input coupling caps C505-506 on the left, and VCA Drive output cap C530 on the right which was in backward.

BEFORE




New Nich ES caps, and the FG cap for C530 was put in backward like the original.. because the voltage across it was 0 to -1.4V.

AFTER



A Nich FG cap was used because it fit under the nearby inline plug, and they had to be used-up.

Over-the-edge view of VCA Drive and INPUT coupling caps with one each of their lead soldered. Everything has to be lined up just right!




Double sided board

It's important to pay attention to EVERY detail.

I wasn't aware that one of the front-most COMPANDOR relay caps C531-532 was a double-sided job.




C531-532 desoldered and removed.. trace on topside highlighted.



Luckily nothing got damaged.


THIS TOPIC CONTINUED in Post #116 below.


-Greg

 

[dubai2000]

Not to hijack this thread, but I ventured today to remove the 2.2µF input caps - and it confirms the old saying: the best caps in terms of transparency is....no cap. My final go will be replacing the carbon resistors on those input boards as well - after that the MA 6200 amp will be ready to sing for the next couple of years!

Wolfram

 

[Bubo]

Gregory,

Is there any consistency to the orientation of the caps on the boards, the absence of a silk screen accounts for a lot of the mis-oriented caps?

 

[jlovda]

To be corrected.

 

[Gregory]

Gregory, Is there any consistency to the orientation of the caps on the boards, the absence of a silk screen accounts for a lot of the mis-oriented caps?

Hard to say really.. the caps are generally oriented the +s in one direction, but not always.

There aren't any silk screen or 'transparent' views of the C33 PCBs like in the MR80 Service Manual. There's a difference between a Service Information booklet, and a Service Manual, i.e the tuner. I think there might have been changes or Service Bulletins on some of the boards. The 1uF `voltage rail bypass cap´ on the LOUDNESS board, I'm almost certain was in backwards, but the COMPANDOR caps were less obvious. They (the compandor caps) had a volt and a half or less on them. Maybe someone will chime in if there's a clear problem.

Hey.. am glad that it's working as expected and very few things got broken.


As for the op-amps, will be listening for a while before rolling those. Probably have replaced perfectly good part already!


-Greg

 

[Gregory]

Nice work. You didn't show a "10,000 ft" view of your work. Did you remove the main board from the chassis? Getting to the area next to the pushbutton switches is not easy.

Thanks John.. know exactly what you mean by a 10k ft view. The web servers have been acting up, so it might take a while to post a high-Alt view.

I've never removed the main board from the chassis, nor the front panel. Everything except the EQUALIZER PCB can be recapped w/o major disassembly AFAIK. That little FG tucked under the front was a bear! Those two caps are 10uF on the lines to the COMPANDOR Record/Listen LED lines.


There was a `white balance´ issue with the photos this time. I used a high-intensity desk lamp with G8 type 25W lamp, so it doesn't look like outdoors.


-Greg

 

[Gregory]

Compandor section recap – continued from Post #110 above.

Luckily nothing got damaged.

Here's how I installed the replacement.



New caps in place with double-sided one elevated. Not too unsightly and I believe they're soldered in ok.


These two caps were replaces with FG because they were on-hand. I think Nich VR would have been fine for the relay circuits.




Muting Control/Delay

Pretty sure this FG was the LAST remaining cap on the main board to get replaced. A type VR would have worked.




Originals

These are the original 10 COMPANDOR caps, and 1 Muting cap.

Not only does the rubber bench mat protect the equipment, it keeps things from rolling around.




Wrap-up

Very intense day trying to get the last few caps replaced. There's still the EQUALIZER board caps, but the front panel must come off to get at them. There are 4 more caps on the Monitor Amp final output stage to be replaced.

The C33 is back in the rack, and the COMPANDOR is working as expected. No problems have been detected due to updating, servicing or recapping this year. The remaining Update tasks should be completed by year-end 2013. Then the C33 Update Project 2013 will be a wrap!


The C33 MONITOR AMP is rated at 0.01% distortion whereas the MC2205 is rated at 0.1%. Let's put it this way.. we'll be listening through the Monitor Amp for a while.


Yes.. it does sound very good.


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Original Post #116 – Forensics anyone?

Using spent ziplocks to stash the original C33 caps.

Caps from January 2013 are on the LEFT (next to empty zip) and more recent ones to the RIGHT.


At the risk of going overboard with photos.. let's have a look at 64 caps changed this year.


January 2013 caps are from High Level Input, Top Cover and Main boards. Center zip is incorrectly labeled.. should be LOUDNESS & PS.


A few things to point out

* Two original 100uF Power Supply caps on the far left (yellow arrow.) Those were replaced with FGs in January (yellow arrow in center) and finally upgraded to big black KZs last week,

* There are 24 NP (or BP - thanks for clarifying John) caps in dark blue and gray. Replacing these should have a direct affect on sound quality, not considering noise floor that is,

* There are STILL four Monitor Amp caps to change, and the EQUALIZER board hasn't been touched yet. Should get the Mon Amp done by the end of 2013!


Think they should be thrown out? Old caps don't take up much space, and all old parts (there are exceptions) are stored for reference.


-Greg

 

[Gregory]

More than a year has gone by since the C33 project started

The built-in monitor amp has been powering the system since the preamp was put back in service. There hasn't been any sign of trouble, I'm glad to report! After a two month `exercise´ of the C33 and its 20 WPC amplifier section, I see this morning that FOUR caps still have to be replaced in the final output section.. pointed out in RED below.




Still pondering which caps to use. KZs would be best, but will they fit in the tight spaces? FGs would fit for sure. Will need to order some of each to check 'em out.


Hey.. this might be a good time to do some Op-Amp rolling.


-Greg

 

[jlovda]

Ironically, in one of my C34V pre's, one of the 150 ohm headphone resistors, R1052, bulged and split in the middle. Luckily it's an easy part to replace. The service manual specs them at 1 watt but all four in mine were 1/2 watt.

 

[Bubo]

Hey.. this might be a good time to do some Op-Amp rolling. :scratch2:



-Gregory

I probably don't need to ask, but are you able to measure the difference in SN and distortion with the op-amps?

Once the sockets are installed, this could be a whole new pastime.

 

[rpoli2518]

Hello everyone,
I was doing my project for C33 so good now its quiet. Doesn't power on. When I did power on for the first time after recap, smell came from PC board 045809. 10uf cap
C1302 was blown, AGC 1A 250V fuse was blown, resistor 820Ohm was very hot. I checked and found the connection between 10uf bipolar cap pins were shorted during soldering new cap. Fixed soldering mess, replaced new resistor 820ohm, new cap
C1302, new bipolar cap, new diodes 1N5395, new thyristor Q1301 which is Omni C106MG. Powered on and again nothing, but hot resistor 820ohm and smell from this board.
What else it might be? Need your help please.