View Full Version : MMKs in feedback

James Evans
14-06-07, 02:56 AM
Fairly old ground I know, but it's taken me some time to get round to acquiring these sizeable little boxes.

Firstly, must commend Elfa on their customer service - ordered 4 caps last friday and they arrived from sweden on wednesday - speedier service than some UK based companies I've dealt with.

So, here they are, wife goes out to yoga wed eves and the nipper amazingly decides to have an early night. Can't believe my luck...

Whip the OA21 out of the rack and dive inside (must say its a damn site easier tinkering inside this than the claymore and obelisk3). Out come the 2 100uf 6.3v caps, and after some wedging, in go the mmks. As a full update, the 4.7uf power amp coupling caps have been replaced with 3.3uf polyprops and psu caps are mundorf mlytics, which I managed to squeeze into the existing pcb slots in the end. Pre section is passive in the onix so all I've now got between the output opamps (lm4562) in my rcd1070 are 10uf polyprop coupling caps in the cdp, then the polys and mmks in the onix - no lytics anywhere, bonanza.

It's been playing for an hour or so now running through some standard test tracks and initial impressions are very positive - no loss of bass at all and things have tightened up down there. Much much cleaner highs - instrument and voice separation is vastly improved. High hats sound more realistic, speedy snare jiggery pokery on cinematic orchestra's everyday is crystal clear - every beat can be distinguished. Harshness reduced as well - can go louder without fatigue.

Happy camper thus far.

I've got 2 more, so the claymore goes under the knife next to remove the 100uf cerafines currently on feedback duty.

Then its off to order 4 more for the obelisk....


and inspired by PDs post on his power amp refurb I had to try out she's lost control - there is indeed a sub bass part that was previously indistinguishable, marvelous, now, off for more listening...

14-06-07, 04:02 AM
Glad to hear everything worked out for you with the MMKs James! They are just perfect in there aren't they. On tuesday I am comparing my oscon decoupled mmk in feedback equipped 102 to a friends 82, will report back with that.

Cheers Magnus

P.s. We will also compare my MiniCapsc to his standard HiCap. d.s.

14-06-07, 05:25 AM
Hi James, Have you got a schematic for the Onix? Had a quick look in mine the other day with a view to doing a few mods but changed my mind because i didnt want to bugger it up.

14-06-07, 05:32 AM
. . . Also, are the feedback caps easy to find?

James Evans
14-06-07, 05:50 AM
Cheers Magnus!

Hi FM, I don't have a schematic for the onix - I read somewhere that it's based on the same hitachi design as the claymore was so made some assumptions based on that... There's only 4 electrolytics in the power amp - the 2 coupling caps and the feedback caps (if my assumptions are correct).

The linked to image shows a photo with them highlighted - the red arrow and circle are the 4.7uf input coupling caps (I know this for sure as I traced the pcb tracks back...)

The two small caps highlighted with the green arrows are the feedback caps if my assumptions are correct.


Hope that helps!

14-06-07, 05:55 AM
As it happens, the Maclaren is also based on the same Hitachi application note; it just use 3 pairs of output devices (can do 24amps continuous!) and a bigger power supply, with regulation to the driver board. The Maclaren also has tweaks to the compensation and Zobel network to improve stability with reactive loads.

James Evans
14-06-07, 06:52 AM
Sounds like a beefier amp than any I have here then. If these are anything to go by I bet it sounds rather good!

While I'm on the subject of replacing electrolytic caps, below is the obelisk power amp schematic. Other than the input and output coupling caps, which would be worth replacing? I'm thinking C5 (100uf/25V) is a definite, but any others?




14-06-07, 07:29 AM
C18 is the bugbear - you have to use a 'lytic, but the whole signal passes through it, and all the nonlinear effects are present big time. Use the best, fresh one you can find, but the single rail circuit holds performance back horribly.

Marginally worth swapping C5 and C1 for films, but the circuit is so nasty that you are basically gold-plating a turd.

14-06-07, 08:52 AM
James, Your a star! Will have a bash at it after my tea. Ive got some MMK's for feedback but im not sure what to use for coupling. I like BC128's in my Naim kit but i have got some rather large 10uF polypops (Mundorf Mcap MKP) i could try. . . Cheers and sorry to hear your other amp is a turd!

James Evans
14-06-07, 08:57 AM
I know, bit harsh that... ;)

It actually sounds ok, although is resigned to living room/tv duty.

You may find you'll have to remove the soap link to be able to fit everything in - I did - not a problem as I don't have one anyway... The polyprops may be a bit big as well as the MMKs... BC128s are well worth a go and will fit in the available space allowing easier fitting of the MMKs.

Enjoy your tea!

14-06-07, 10:18 AM
Thats funny, my turd's not even gold-plated!

James Evans
15-06-07, 03:38 AM
Hi Andrew, managed to completely blow up your obelisk yet?

15-06-07, 05:28 AM
I fitted the poly's in the end because the BC128's are only 25v, the original caps were 63v. Any idea what the voltage measures in this position?. . . Its hard for me to comment on the sound because i only use the amp for the tv but will compare it to my modded Naim kit sometime.

James Evans
15-06-07, 05:47 AM
you should be fine using the bc128s as coupling caps I would have thought, having said that, the polys are probably superior to the 128s anyway...

22-07-07, 01:27 PM
Bought some BHC T-net caps from patrick dixon and fitted them saturday. First impressions were not good, with bugger all bass and splashy treble. Had a listen today - good cd's sound better (more detail, smoother and better bass resolution). A cd that i thought was a bad recording (thin/lightweight) actually sounds rather good now. They're a pain in the arse to fit but well worth the effort.

James Evans
03-08-07, 03:34 AM
took out the 3.3uf polyprops I was using as coupling and put in some bc128s and after a few days I've got a lot more bounce back, without losing the "air". Got MMKs in 2 amps now and they're sounding great.

03-08-07, 05:54 AM
Interesting! I hav'nt tried BC128's yet. You should try bypassing it with a 100nF polypropylene, they work well with BC128's - they give smoother, more detailed high frequencies. My Cambridge audio 640c has, what looks like a 100nF polyester cap, bypassing the electrolytic coupling cap, as standard.

James Evans
03-08-07, 06:02 AM
cunning. I've got a couple of 100nf polyprops here so I'll give that a go when I get the chance. It's a bit fiddly soldering under the pcb in that position as the L-bend of the backplate overhangs that part of the pcb - will play around though.

It's still not got the dynamic punch of my claymore, but has a lot more open feel to the top end, so for the moment I'll just enjoy listening to one for a few months and then swapping back!

James Evans
07-08-07, 02:09 PM
right then, had a bit of time to myself tonight, so decided to drop in these 0.1uf polyprops. They were too chunky to fit under the pcb, so had to take out the bc128s, put the polys in their place and then whack the 128s under the pcb, which was a damn fiddly tight squeeze as well.

Only just finished and first impressions are.... it all still works.... which is a bonus

listening to be done tomorrow...

08-08-07, 08:59 AM
Come on James! Whats the verdict? Are you happily listening to music or cursing me?

James Evans
08-08-07, 09:21 AM
well, been listening most of the day now. There's definitely something different about the HF presentation. Can't quite put my finger on it, but I've been hearing little details I've never heard before, such as a rasp on a "sp" sound in a vocal, more detail to a bass note, that kind of thing. It feels like I've gained the openness of the previous 3.3uf polys, but retained the warmth in the bass of the bc128s, and without any of the ear-splitting treble. Worth doing I reckon, even if it is a tad fiddly. At the moment its one of my more successful bypassing capers - always removed them in the past...

08-08-07, 09:54 AM
James, good stuff.

My Rega Planet used to have Ceramic 100nf bypasses everywhere, all the ICs, the lot, I changed them out slowly, one by one for metalised polypropalene and had a listen and a settle down period between each, even in the digital sections (supply decoupling, not signal) the effect was unmistakable!

Took it even further last week, by sheer fluke, I happened upon a bag of what must be the last 125v 100nf axial polystyrenes on the planet (made by Ducati) I knew what was coming, and how!

All the 'styrenes in the bag went in, around fifteen or so, one by one again with a listen in between, lovely, just top upgrade, so much goodness, and from CD!

Gobsmacked by that result, I wonder what bypassing the 100nfs with 100pfs would do..so I did....

Polystyrenes at that size maybe as rare as hens teeth but I'm betting that film/foil polyprop FKP at 100nf types won't be far behind.

Edit: I replaced COG ceramic types and ceramic discs directly with foil polystyrenes not metalized polyprop, the polyprops replaced the supply bypass ceramic beads originally, I'm not sure that metalised polyprop types would be an improvement over the COGs and discs.

08-08-07, 11:31 AM
Nice one James! Looks like im going to have to get the soldering iron out. . . Mark, Im suprised you noticed an improvement by changing the ceramics in the digital section. I thought i read somewhere that they were the best capacitor for that job. Could have me wires crossed though!

08-08-07, 01:33 PM

I think cost and wave soldering are the reasons they don't use them, I could also be wrong but the effect was definately right though!

James Evans
09-08-07, 12:59 AM
I was tempted to try polystyrene myself - jab electronics have 47nf (http://www.jabdog.com/cap-sty.htm). Figured I may as well try the polyprops first though as I have them here...

09-08-07, 02:11 AM
I tried polystyrene bypass/coupling caps (100nF) with BC128's in my Naim kit and didnt think they were as good as polypropylene.

09-08-07, 08:34 AM

If the 'styrenes you used were either 30v or 63v then I'm not suprised, I tried these two ratings (Sufflex) and the props were better, the 125v Ducatis though were better than the 250v props and I have some 630v rated 10-22N rated styrenes yet to try, only problem is the size though, they are very large, add to that a mild modders burn-out a left over from my kit being in a constant state of flux, that will stop when I get another player to mod that will stay outside the main system.

The things we do eh?

09-08-07, 09:05 AM


Jeckel: I can't be arsed.

Hyde: ah go on!


I have a couple each of these going spare for experiment, 1000 and 1500v 47n and 10n polyprop film/foil, gratis, give them a go and see what happens (if anything).

09-08-07, 09:54 AM
Just had a look at the 'styrenes and they are 10nF not 100nF but must be high voltage because the legs are much fatter than the 160v ones. . .

09-08-07, 10:00 AM
Fitted the BC128's last night - sounds pretty good on one channel. . . Oh dear, time for a new thread :(