Kenwood KA-7100 recap-refurb

Status
Not open for further replies.
I picked up a non-working KA-7100 a few months ago on eBay for a good price. The previous owner had correctly diagnosed speaker relay issues, but did not know how to repair the amp. Upon searching old threads, I found several references by EchoWars about how this amp is prone to relay failure due to aging components. So I felt it likely that I would be able to repair the amp simply by replacing components in the relay circuit, which indeed was the case. I believe this amp has been a workhorse judging by the way she looks inside and out.

As I worked through this project, my approach was to study existing threads about the KA-7100 in order to have an idea of common areas of failure. The issue mentioned most frequently in old threads is the speaker protection relay circuitry as noted above. Another area of concern is the regulated power supply, both the components on the main PCB as well those serving the control board.

In addition to addressing common areas of concern, I made component upgrades and a couple of conservative modifications. Also, a couple of observations (in regards to the power switch and PCB supports) are included in this thread.

I hope that this thread can be helpful to others working on a KA-7100



Overview of work done:
-power supply received special attention in refreshing and beefing-up of components
-heatsinks added to regulator series pass transistors
-AC power switch's unused contacts employed, new safety cap installed
-electrolytic and tantalum caps all refreshed and/or upgraded with film
-previous repair (by prior owner) of Qe10 redone with a better sub
-new Bournes trim pots
-new speaker relay, relay driver transistor, and diode
-output transistors received new mica and grease
-several stressed-looking resistors replaced
-all zener diodes replaced (power supply, main amp, control and phono boards)
-missing foam PCB supports replaced
-DeoxIT D100L used to clean pots, switches, fuses/holders, RCA's, speaker terminals, face plate, knobs, switch covers
-bias and DC offset adjusted to spec


Acknowledgements and thanks:

-EchoWars: numerous previous posts/threads that serve as valuable and useful resources. One in particular regarding KA-7100 power supply will be linked in this thread where appropriate.
-leesonic for this thread: Another Kenwood KA-7100 recap which, for anyone recapping a KA-7100, can be used as a general guide with great pics and his complete parts list included.
-ManyMoonsAudio for assistance in a separate thread where I had questions about the AC power switch. See post 14 below.
 
Last edited:
power supply before and stripped


power supply stripped and cleaned
EDIT: In this pic you can see where I "stretched" the holes for the leads of the new main power supply filter caps. I did this using an X-acto type knife (Americaline brand actually). For anyone who may try this please wear eye protection because the tip of the blade could easily break off under pressure.


1)power supply before - wide view (from rear of amp)
2)power supply stripped and cleaned

01 DSCN0577 resized 232kb .JPG

02 DSCN0585 resized 237kb.JPG
 
Last edited:
power supply re-populated

Of note in this pic are the UF5402 rectifiers and the new Omron MY4-02-DC24 relay. Diode D27 (part of relay circuit) just out of view was also replaced with 1n4003.

03 DSCN0200 resized 234KB .JPG
 
Last edited:
Nice clean work. The one I restored didn't need quite that much work done...relay was working properly, etc., so I left it alone. Thread here.
 
regulator components run a little hot

When beginning a project like this, it has become my SOP to assess any areas where heat might possibly be an issue. According to information found in other threads, and confirmed in my sample, the one area in KA-7100 where heat may be a concern is in some of the components in the regulator (low voltage rails feeding the control and phono boards).

I am not saying that the heat there is outrageous, I am sure there are many other designs with more serious heat issues. To be fair, this amp was still working fine (except for the relay) more than 35 years after it was made.

But a couple of the resistors were looking a bit over-worked. And the series pass regulators were putting out a respectable (and continuous) amount of heat. This R77 tested OK out of circuit, but I have read one recent report where it had failed.


R77 before pic
index.php
 

Attachments

  • resistor  R77 hot DSCN0549.JPG
    resistor R77 hot DSCN0549.JPG
    124.4 KB · Views: 1,531
Last edited:
regulator rebuild

I wanted to give special attention to the power supply, and the warm-to-hot running regulator in particular. So I was considering ways to beef-up this area. I found this thread where EW gives instruction on KA-7100 regulator rebuilding in posts 26 and 28. Recommended were using MJE15032/MJE15033 for the series pass transistors, using series'd 14v zeners to replace the 28v D17/D18, and increasing power rating of R75/R76/R77/R78 from .25w to .5w.

Zener diodes in positions D17 and D18 each received two Vishay TZX14B in series. These are low noise, 14V, 2% tolerance zeners. With these in place, voltage drop across D17 and D18 read around +28.5 and -28.5. This yielded final regulator output voltages of close to +28V and -28V, which is what the schematic shows.

This is a revision from my original rebuild where I had used two TZX14C in series which yielded final regulator output voltages of around ±28.8V. That probably would have been OK, but I wanted to try to get it closer to the value shown on the schematic, and changing to TZX14B made that happen.


I went over-kill on resistors R75/R76/R77/R78 and used some very pretty Vishay-Dale 2w, 1%, CPF series, metal films (same as leesonic in the thread linked in my OP)

Capacitors in the regulators are Panasonic FM. By chance, the jackets on these have a color scheme very similar to the Nichicon KW filter caps. But FM is what I wanted to use there, and I did not choose based on their color, lol.

05 TO-220 - DSCN0231.JPG
 
Last edited:
heatsinks added to regulator series pass transistors

leesonic did this on his KA-7100 and I liked the idea as a way to assist in moving heat off of the transistors and out of the area. Not necessary, especially with the more robust transistors installed, but a cool added touch (pun intended, I guess).

I used a different, more shallow, heatsink than lee because the series'd diodes in my amp are up in the air, and closer to the transistors than the original D17/D18 which are mounted close to the PCB.

Mouser part # for the heatsink: 532-577002B00. I also used Aavid Thermalloy TO-220 mounting kits (Mouser 532-4880) and liked them a lot. Very nice the way the nut tightens down when using the fiber washer. I put Permatex Blue (medium strength) Threadlocker on the screw threads to help prevent loosening in the future (you can't see it in the pics because it turns white/clear when dry).

I did not use the mica insulator that came with the TO-220 mounting kit, opting instead for just a touch of standard heatsink grease spread very thin. Prior to mounting the heatsink I gently cleaned surface oxidation off the backs of the transistors.



These little heatsinks made a greater difference than I would have thought. Celsius temperature readings at the front tab of the TO-220's with and without heatsinks, after running 1 hour at idle:

_Q31_____Q32

58.7°C___63.3°C ____<<<without heatsinks
49.9°C___52.7°C ____<<<with heatsinks

06 heatsinks DSCN0236.JPG
 
Last edited:
upsizing main filter caps to 10,000uF

This is, of course, one of the first "mods" that re-builders like to jump on. The more responsible among us warn that added capacitance in the main filter caps results in greater inrush current which can stress rectifiers and the AC power switch (sometimes the main transformer is mentioned as well).

I went ahead and ordered 10,000uF Nichicon KW's anyway. But any reservations that I may have had about this disappeared when I realized that the KA-8100, which uses 10k filters, also uses the exact same rectifiers and AC power switch as the KA-7100.

<EDIT: EW makes a good point in the next post. I actually opened up the power switch and inside observed some pitting on the contacts as well as a carbon-like film. Having seen that, I would definitely NOT recommend upsizing the filter caps unless you also give the power switch some attention. As it turns out though, there is a reasonably simple solution as will be described shortly>

Originally, before I discovered the shared components mentioned above, I chose UF5402 rectifiers mainly because of their higher Ifsm at 150A vs the original GP25D rectifiers at 100A. Ifsm is the spec that quantifies the short term current capability of a diode. A possible added bonus of the UF5402 rectifiers, as stated in the Vishay spec sheet, is their "Soft Ultrafast Recovery". The verdict is still out (hotly debated over at DIYaudio) as to whether that is a true benefit, but surely it can't hurt.

Lead spacing on the KW's does not quite match the holes in the PCB. But just a tiny bit of careful "stretching" of the holes with a small file or sharp blade will get them to fit nicely.


<EDIT August 2020: If doing this project today I would use 63V rated main filter caps instead of 50V. The rail voltage is around 47VDC if I recall correctly, and I would like to have a bit more cushion. That and, AC supply grid voltages are slowly increasing. Source AC voltage from the wall affects the non-regulated main rail voltages.>



DSCN0581.JPG
 
Last edited:
Those toggle power switches are a rare and vulnerable component. Added capacitance can exacerbate the situation.

http://www.audiokarma.org/forums/showthread.php?t=504673

I have read that thread. Great idea there. I understand that this mod would be a relatively easy way to make an AC power switch last indefinitely. (easier than installing a separate power relay)

For those who have not read it I will attempt a summary: a triac is used to route the bulk of the AC current around the original switch, leaving only a minute current through the switch, just enough to keep the triac operating.

After reading the thread (a couple times) I still have questions. I would prefer to ask those questions in that thread where they will be more relevant and will be seen by more people. I do have some projects in the queue that might be good candidates for this mod.
 
AC power switch

I had never looked inside the power switch of a project amp. But since I did upsize the filter caps in this one, and it is "vintage", I decided that it would be a good idea to open this one, just to see if things looked to be in good shape.

I am very glad that I looked inside because I found two things that I did not expect:
1) there WAS some pitting of the contacts as well as a black carbon-like film
2) there were four pairs of contacts in the switch, only ONE of which was being used. The other three pairs of contacts were clean and new, and were not even wired to anything.

See this satellite thread: Question about the power switch used in KA-7100, KA-8100, KA-9100

The same switch was also used in KA-8100 and KA-9100, as least in the USA and Canada. And in KA-8100 only one set of contacts was employed just like KA-7100. (don't know about KA-9100).

The fresh, unused contact pairs gave me a couple options. First, by simply moving the wires and safety cap from one side of the switch to the other I would have access to an unused pair of contacts. However, even better, by disassembling the switch and flipping the rear part over, then the bottom of the switch would have two fresh pairs of contacts that could be paralleled.

The second option is what I chose to do. Essentially there is a new switch in place with twice the contact area than was utilized by the original (in KA-8100 as well as this amp).

Getting this switch apart and, especially, back together is a challenge not to be taken lightly. While on this subject, the mod linked above by EW would probably be a lot easier to implement (and a better long term solution). IMHO unless someone is willing to either parallel the contacts as I did, or employ the triac mod, then it would better not to upsize the filter caps.

Note: Unfortunately, these pics do not show the black film that was on the contacts because I cleaned it off with Deoxit before I thought about taking pics.




1) stationary contacts in original orientation (lower right shows damage)
2) rocker contacts, showing pitting on one (lower left) and three undamaged
3) final assembly with contacts (brown part in rear of switch) rotated 180°, jumpers to parallel the contacts, and new safety cap

contacts.jpg rockers.jpg final.JPG
 
Last edited:
Your attention to detail is admirable! I am not surprised you average a year per project.. I can see hours on end of work gone on there. nice job
 
speaker protection relay circuits

As noted in the opening post, the relay in this amp was not functioning. Taking readings from original components of the relay circuit revealed two that were in extremely bad condition.

C41. In fact, of all the original caps that I have removed and measured (every cap from 5 project amps so far including this one), I haven't seen anything close to the condition of C41. This was once a 10uF cap. Using my BK Precision 878 LCR meter I had readings of .2uF with a DF of 1.7 Both of those values are seriously bad. The uF reading speaks for itself. For perspective, the DF of the new cap replacing C41 was .035 (lower values are better for DF, similar to ESR).

Q34, the relay driver transistor, was also in bad shape. I got an hFE reading of 33 which, originally, should have been in the range of 100-200 for a 2sc1213A(C).

It is no wonder the relay would not close. C39 and C40 both measured OK. Everything was replaced of course.

Q34 > KSD1616A(Y)
C39 > Nichicon VP (bipolar)
C40 > Nichicon KL (low leakage)
C41 > Nichicon PW (105°C)

original Q34 and C41. heat from the transistor and its close proximity may have accelerated the demise of C41 so I used a 105°C cap here

DSCN0547.JPG
 
Last edited:
Your attention to detail is admirable! I am not surprised you average a year per project.. I can see hours on end of work gone on there. nice job


Thank you RuffzGuts. Well I guess this project only took about 9 months. But I have another started before this one that is still opened up :sigh:. Of course I am not working on them every day.
 
main amp board

Except for the relay circuit caps mentioned previously, replacement caps on this board are all Elna Silmic II. There were no significant heat sources near any of these except the relay driver transistor Q34 (see post #17).

All trimmer pots were replaced with Bournes 500 ohm multi-turn. Mouser # 3296Y-1-501LF

All zener diodes on this board were refreshed. De1/De2 got BZX79C24. The 28v De3/De4 got series'd 1N5244B. <EDIT: I am going to revisit De3/De4 to measure current and see if possibly a single zener might be a better match. I had no 28v zeners at the time I replaced these so used 2x 14v out of convenience only>

Qe10, originally 2sa915, had been replaced by someone with 2sa850. Probably not the worst possible sub assuming someone pulled it out of their stock on hand. But it is a .8w part whereas 2sa915 is 1w, and some of the other parameters also did not precisely match.

I was able to find 2sa916 which is nearly identical to 2sa915 except that it has higher Vcbo and Vceo than the original part. All other parameters match exactly. I went ahead and replaced Qe7 through Qe10 with new 2sa916 parts and matched them all for hFE (matching hFE for these locations might not have been necessary, I don't know. Just an OCD thing).

!DSCN0217.JPG
 
Last edited:
Very nice work, and thanks for the link to my thread, I'm glad it was of help to you.

Lee.
 
Status
Not open for further replies.
Back
Top Bottom