Hi,
I'm starting on a recapping project on my Meridian 557. I am a EE/CS and have done lots of systems work from prototypes to production commercial control systems, computers, etc. However, I haven't done that much work on hifi amplifiers and I've never tried to remove an amplifier board from a 5 pound heat sink.
On the 557, the main circuit board (similar to the Alex MM circuit board in "The Very Best Amplifier I Have Ever Heard" thread) has 4 power transistors (Toshiba 2SA1302) on the top and 4 on the bottom configured to lay flat on a heat sink that the board is attached to. Each of these power devices has a very thin custom square of heat tape attaching it to the heat sink along with a single hex screw. My main question is whether anyone can provide the best process for removing and reattaching these power transistors to the heat sink so I can flip it over to remove/replace the caps. There are other devices on the heat sink, some with heat tape, some with thermal compound but the power devices are the only problematic ones.
Thanks,
D
I'm starting on a recapping project on my Meridian 557. I am a EE/CS and have done lots of systems work from prototypes to production commercial control systems, computers, etc. However, I haven't done that much work on hifi amplifiers and I've never tried to remove an amplifier board from a 5 pound heat sink.
On the 557, the main circuit board (similar to the Alex MM circuit board in "The Very Best Amplifier I Have Ever Heard" thread) has 4 power transistors (Toshiba 2SA1302) on the top and 4 on the bottom configured to lay flat on a heat sink that the board is attached to. Each of these power devices has a very thin custom square of heat tape attaching it to the heat sink along with a single hex screw. My main question is whether anyone can provide the best process for removing and reattaching these power transistors to the heat sink so I can flip it over to remove/replace the caps. There are other devices on the heat sink, some with heat tape, some with thermal compound but the power devices are the only problematic ones.
Thanks,
D
Ok, after some simple experiments I've found the trick for removing the transistors from the thermal pads with zero damage to either. Whatever you do don't try to pry them up, I learned this lesson years ago. After you've removed the hex screw holding them down, get a very small, very sharp flat blade screwdriver like the ones you might use to fix a pair of glasses. Fit the blade between the transistors and slowly wiggle it back and forth. Don't use much force, the trick is accumulated motion because once it moves 1/1000th of an inch you've won and it will come loose.
If you end up with one solitary transistor then find a phillips head screwdriver that fits the mounting hole as perfectly as possible then use that to wiggle it gently back and forth being careful not to touch the threads for the mounting screw.
If you end up with one solitary transistor then find a phillips head screwdriver that fits the mounting hole as perfectly as possible then use that to wiggle it gently back and forth being careful not to touch the threads for the mounting screw.
Last edited:
Well, the first 1000uF 16v cap I pulled out measured 1 uF and the second one was 450 uF. I guess this was more than due for a recapping. Meridian used Nichicon Muse for the 1000uF 16v but I'm replacing them with Nichicon Fine Gold. Both the Muse and Fine Gold caps are significantly larger than the ones made back in the mid-late 1990's when these amps were made but the larger ones will fit in all locations if you are careful. The Nichicon 10uF 50v caps are the same size as the old ones so no issues there.
But OMG, what a difference these new caps made. It's like having a new amp. The clarity, staging, bass and high-end are all improved. Well worth a day's work.
D
But OMG, what a difference these new caps made. It's like having a new amp. The clarity, staging, bass and high-end are all improved. Well worth a day's work.
D
One final follow-up on this project. I measured every cap as they were removed and all were within spec except for C3, C18 and C43 which were all Nichicon 1000uF 16v caps and those were wildly out of spec. C3 which filters the -15v PS was completely fried on both channels measuring 1uF on the left channel and 143uF on the right. The other 1000uF 16v caps measured between 274 and 492 uF. The schematic refers to C43 as a 1000uF 50v cap but it was a 1000uF 16v like C3 and C18. I'm assuming the assembly team had it right since C43 is filtering the + input of the op amp driving the bias circuit where the fine adjustment from VR1 comes in which is balanced between +15 and -15 so you can assume it's very low voltage there.
Glad the OEM sound is back.
I've put over 400 industrial grade nichicon, rubicon, panasonic, and occasional United chemicon or multicomp where I couldn't buy the previous in right size or grade. Not the 500 hour cheap ones, the >3000 hour service life ones. Some of 10000 hour service life rating or higher if I could get it. So far there is much improved sound or function in my organs, and more power or volume to the amps mixers equalizers etc. I'm not at all convinced of the value of the "audio" grade caps, which have a huge price premium. I buy mine at regular distributors like newark & digikey, who publish the service life spec in the selector table if I ask for it. Digikey has been known to sell 8 year old caps for full price, so definitely reform before use. Newark/farnell puts a discount star next to the line number when the shelf life termination date is getting close, & cuts the price. Shelf life has to do with suitability for pick & place machines, then hammering the cap with full voltage first power up.
I guess people are going to sell things for whatever people are foolish enough to pay. I got the 1000uF 16v caps on Amazon for just over $2 each and the 10uF 50v caps were about $1. The big 10,000uF 80wv caps were about $7.50 each. I couldn't believe there were a couple companies selling the big Nichicon caps for over $100 each. I'm going to guess that the $7.50 caps sound the same as the $100 ones. ;-)
In January I bought 1000 uf 35 v Panasonic 10000 hr 105C caps for $1.06 each. Newark. In April I bought 10 uf 50 v 5000 hr 105C caps for $.16 at digikey. The 4700 @ 63 v 13000 hr @ 85C kemets were $6.54 digikey. I've had no trouble with stock like these. When I was buying caps from the front shelves of the TV parts stores in the seventies, like sprague atomlytic (no service life spec, I have the datasheet), they were lasting about 6-8 years @ 2000 hours a year in my ST70. 4 times I've re-e-capped that amp. You see my passion for XXXXX service life hours caps. Nothing has been replaced that I installed since 2008.
Happy listening. I've got NY Philharmonic orch. on radio tonight, through a 1990 monaural amp re-e-capped last year.
Happy listening. I've got NY Philharmonic orch. on radio tonight, through a 1990 monaural amp re-e-capped last year.
Elna Tonerex and Nichicon Fine Gold audio grade are not too expensive in the typical quantity need to recap an amplifier, usually well under AU$100 worth. They do measure more linear / lower loss than 'standard' electrolytics. I've never tried anything more expensive. I test the capacitors I use for Vloss, ESR, value, and dissipation factor just in case some fakes are slipped in, but I haven't found any duds yet.
I recently recapped a Marantz PM-64 Mk II, which anecdotally isn't as good as it predecessor. I bought a Fine Gold recap kit from Audio High End. As per usual some values were mixed up in the kit and there were some spare caps (they seem to have put in values for the PM-54 as well which uses the same service manual).
But the revelation was replacing three small value electrolytic capacitors, 1µF or less, in the 64's audio chain. Replacing these with WIMA film caps (not supplied in the kit) made a huge difference. Price could not have been the reason why Marantz used electrolytics, it must have been to make sure that the 64 Mk II didn't sound as good as its more expensive brethren.
Thanks for the info. Since my primary job is doing large scale Internet software systems, I tend to lose the hardware side of my EE/CS education. Because of this I typically purchase parts that lots of people swear by at the most reasonable prices I can find. I figure the few bucks I might be able to save will always be less than the time I would risk if a lower-priced part failed and had to be replaced again.
I know a lot of people think that happens but I've spent quite a few years working in and managing hardware and software teams nobody in finance or marketing has ever suggested anything like that to me. Engineering teams are always kept focused on quality and reliability since lack of either can hurt a company far worse than cannibalization of the product tiers. However, I suppose it's possible given a stupid enough executive team. ;-) My guess it was more likely just a design mistake.
D
I sure trust Newark & nichicon/rubicon/panasonic service life estimates more than Amazon star ratings. None of the >400 e-caps I've put in organs & amps since 2008 have failed. Plus if Newark caps are nearing shelf life end, they star the line and discount them.
Certainly that's possible; if so if was an unfortunate failure of design integrity oversight.
In my experience servicing consumer high fidelity equipment it is not uncommon to come across shortcuts in the lower range products that could hardly be justified on a cost basis, because the cost savings involved are trivial compared to cost factors in other arbitrary design choices in the same series of product.
Besides, if the cheapest product in the range sounds as good as the most expensive the manufacturer is going to forego some profit. There is a strong financial imperative to make sure this doesn't happen. I should disclose that for a few years I was employed as a developer/designer of consumer and professional audio equipment.
Lol, I don't trust Amazon star ratings for much beyond household cleaning products. I meant that I trust components that people who know what they are talking about swear by. However, you can often pick those components up for less at Amazon than the dedicated parts sellers. But on the other hand, Amazon doesn't carry much electronics parts beyond what's popular so the sellers like Digikey and Newark still get plenty of business. I get most of my info regarding parts quality in forums like this one that have a lot of people who are knowledgeable in one field or another. I've pick up a lot of hifi info here at diyaudio.com for example.
You probably have a better perspective on this than I do then. The last hardware project I did was for a LEO satellite Internet company where parts cost was a pretty small part of the equation since chasing thousands of little satellites around space to replace parts isn't a good idea. I imagine in the consumer hifi industry it can be pretty cut-throat. However I've done several Internet startups on the software side and typically being venture capital backed, there is more cost-based corner cutting than I like to think about. However, the trick is balancing the cuts so you don't leave your customers staring at 404 pages too often.
Btw, I just got an interesting page on this site telling me I had to wait for a backup. A couple of things about that - I loved the picture, that had to have been a 1970's era data center since the only place you see equipment like that nowadays are in old documentaries. But why would a system designer make users wait for a backup?
Hi,
I am from France (please excuse any errors in my English). I recently joined diyAudio because I need advice regarding a Meridian 557 power amplifier.I have started recapping this amp, which I bought second-hand. I discovered that a part overheated and a resistor (R553) burned out (see picture) and must be replaced. On the attached schematic, I can see that the Zener diode (D50) is listed as 15V, while on my amp, it is an 18V Zener diode. My amp was built in 2001, whereas the schematic is from 1995.Here are my questions:
Thank you in advance for any advice you can provide.
Thierry
I am from France (please excuse any errors in my English). I recently joined diyAudio because I need advice regarding a Meridian 557 power amplifier.I have started recapping this amp, which I bought second-hand. I discovered that a part overheated and a resistor (R553) burned out (see picture) and must be replaced. On the attached schematic, I can see that the Zener diode (D50) is listed as 15V, while on my amp, it is an 18V Zener diode. My amp was built in 2001, whereas the schematic is from 1995.Here are my questions:
- What could have caused the damage to R553? I don't understand what "Short S.S. Cap" means, to be honest.
- In your opinion, which other components in this area might also be damaged?
- Would you suggest replacing the 18V Zener diode with a 15V diode?
Thank you in advance for any advice you can provide.
Thierry
Attachments
Op amp supplies are often + and - v. If you have a - 18 v supply, then certainly do not replace the zener with a 15 v one. The supplies have to be symmetric.
Zeners are high failure parts. I suggest you replace the resistor with one of equivalent or higher wattage, and then check the voltage across the zener with power on. Warning, do not use two hands to probe an amp with the power on. >24 v across your heart can stop it. No rings or metal on hands wrists or neck. Wear safety glasses, parts can explode.
If the zener does not measure nearly 18 v, replace it with one of same value.
Warning Vishay resistors now have a watt rating at 350 F. Also are tiny. The resistors can stand that temperature, but your board will burn. Buy resistors of at least the same length as the ones now installed.
Since freight from the distibutor is $10 and the zener is $.20, I would buy the zener now.
Zeners are high failure parts. I suggest you replace the resistor with one of equivalent or higher wattage, and then check the voltage across the zener with power on. Warning, do not use two hands to probe an amp with the power on. >24 v across your heart can stop it. No rings or metal on hands wrists or neck. Wear safety glasses, parts can explode.
If the zener does not measure nearly 18 v, replace it with one of same value.
Warning Vishay resistors now have a watt rating at 350 F. Also are tiny. The resistors can stand that temperature, but your board will burn. Buy resistors of at least the same length as the ones now installed.
Since freight from the distibutor is $10 and the zener is $.20, I would buy the zener now.
Last edited:
Hi Indianajo,
Thank you for the advice, including the suggestions related to safety. I will do as you say. We still have some shops in Paris where I can easily pick up resistors or Zener diodes. Capacitors are trickier to find, and I have already ordered them.
I will let you know if I am successful with this project.
Thank you for the advice, including the suggestions related to safety. I will do as you say. We still have some shops in Paris where I can easily pick up resistors or Zener diodes. Capacitors are trickier to find, and I have already ordered them.
I will let you know if I am successful with this project.
??
I would like some addresses, I only know R.A.M boulevard Diderot.
I still have a pied-à-terre in Paris and I like to go and wander around the old shops when I go there.
Much of the circuitry is not shown on this diagram but R53 and the Zener are. It looks to be part of some standby/muting/power on/off detect type of circuit.
For the 10k to burn requires significant voltage and the the fact the 15 volt regulator above it is fed via two series 4 watt 680 ohms suggests a highish (DC) voltage is indeed present. So any 10k should be voltage rated as well as rated just in terms of power dissipation. It would be worth measuring the actual voltage present and calculate the dissipation in the resistor.
The Zener is just a cooking grade part and the wattage can be determined from the above calculation from which we can derive worst case dissipation.
I suspect the arrangement of all these power supplies mean this area runs very hot and all electrolytics will be highly suspect in that case.
For the 10k to burn requires significant voltage and the the fact the 15 volt regulator above it is fed via two series 4 watt 680 ohms suggests a highish (DC) voltage is indeed present. So any 10k should be voltage rated as well as rated just in terms of power dissipation. It would be worth measuring the actual voltage present and calculate the dissipation in the resistor.
The Zener is just a cooking grade part and the wattage can be determined from the above calculation from which we can derive worst case dissipation.
I suspect the arrangement of all these power supplies mean this area runs very hot and all electrolytics will be highly suspect in that case.
I should have written "We still have a couple of shops in Paris. Saint Quentin Radio is another one".
