The KA-9100 restoration.

All the main filter caps have been desoldered and pulled so access to the two diode boards has been obtained. Many of the components on these board are getting replaced.

Here is a look at the first of the two diode boards. It houses rectifiers for one of the two supplies and likely the most unreliable part of the amp, the relay driver 2SC1212A.

5uvlvPg.jpg
 
That was the most difficult board in the amp for me - the access,all the stuff crammed at one end, figuring out what's what on the solder side, reinstalling the board w/o mashing a resistor or something, etc.

 
It's definitely not the most enjoyable board to work on. There's quite a few wires on the rear of the board to be waded through in order reach everything. Bringing the iron to the work and taking it out has to be done with precision. Most of the components leads are bent flat to the board which always makes it more of a challenge to remove them. Patience is key.

Here's the current state of the first diode board. The rectifiers and caps have all been removed. Nearly all the resistors have been replaced. All diodes and transistors have been changed. The only remaining stock parts are the ceramic disc capacitors, the mylar cap, and the inductor.

4joV7iJ.jpg
 
The first diode board is complete aside from the main filter caps. The glue was removed from the board at Cz4 and Cz9. All the 0.25W resistors were replaced with 1% 0.5W metal film. The three stock 1W resistors at Rz5, Rz10,and Rz17 were upped to 2W to match the original size. The large 4.7 ohm 3W resistor was upped to 5W. The two 2SC945 transistors at Qz1 and Qz3 were exchanged for KSC945CYTA (center collector like the original part). The relay driver Qz2 2SC1212A was swapped with a KSC2690AYS. Cz8 10uf 50v, was replaced with a Elna Silmic II 10uf 50v (modern, low esr parts in this value are mostly 5mm in diameter, thus a larger 8mm diameter audio cap was chosen here). C4 is a stock 100uf 16V Bi-polar replaced with a 100uf 25V Nichicon ES. C9 is a stock 100uf 25V low leak cap replaced with a 100uf 35V Nichicon KL. The original rectifiers U05C 2.5A 200V were replaced with UF5404 3A 400V.

This board is challenging. First of course there is the main caps needing desoldered just to gain access. Then there's several small pads that would love nothing more than to lift when you try to remove the components with bent over leads on them. Then add to that a minefield of terminal posts and wires obstructing access to the solder pads and tempting a melting of insulation with a small slip of the soldering iron. Not the very worst board I've seen possibly (a full recap of a Pioneer SX-727 tuner board comes to mind), but also not one a novice has any business touching. I'm now convinced there's a reason the KA-9100 has the PITA reputation it has.

Mrl9rsx.jpg
 
Last edited:
AHH ! PITA ... "Patience In This Application" is it.
I've done a ton of these, seems the the relay driver goes bad alot.

John
 
The second diode board contains four U05C rectifiers for the second supply along with five resistors that I'll be replacing. With the exception of the filter caps of course, there are no electrolytic capacitors present on this board.

Zu7mvXa.jpg



Shown is a very handy lead forming tool at work making lead bends for strain relief. All the 1W and larger resistors and each of the rectifiers receive this treatment.

mXHl7he.jpg



The second diode board shown completed below. As before, the four U05C 2.5A 200V rectifiers were replaced with UF5404 3A 400V. All the resistors are replaced with metal film types of similar size. The 4.7 ohm 3W resistor is upped in wattage to 5W. The 1W 10ohm resistor goes to 2W, and the two 56 ohm 2W go to 3W. There's just one 0.25W resistor here and that is replaced with a 1% 0.5W.

KxIpauz.jpg
 
Last edited:
The meter board has two general purpose 47uf 10v electrolytic caps that were replaced with a couple of Panasonic FC 47uf 25V. The eight carbon film 0.25W resistors are replaced with 1% 0.5W metal films. As mentioned earlier pins three and four are vacant due to the removal (for protection) of the meters.

Js6CJ95.jpg


pHdlj6p.gif
 
The four main filter caps of the KA-9100 are solder tab type. It seems the only drop in replacements are Nichicon Gold Tune which cost $25 apiece and are not available through the large suppliers like Mouser and Digikey. Not wanting to spend $100 for four filter caps I took to finding an alternative. When Mouser had some nice Kemet 15,000uf 63V caps with solder tabs I thought I had lucked into a nice find so I quickly ordered some, but when I received them the differences from the originals were soon realized. I knew that the tabs would not be offset like the stock caps, but the spacing difference was quite a surprise. Having already invested in these caps I turned to an idea that I'd been mulling over in previous builds.

What I'd planned to do was to take a lid from a Powerade bottle and epoxy it atop the cap as a spacer. This basically serves three purposes. It enhances the height of the cap about 10mm, provides a stable base for them to rest, and most importantly allows for the tab spacing to be dealt with via flexible wiring.

Using 22 AWG wire, I wound it once and fitted it into the small hole of each tab in the new Kemet cap. Once through the hole, I wound the wire again to provide a nice 1.5" to 2" lead and soldered them on. Using some large pliers I flattened the leads to allow them to more easily fit through the slots in the PCB where they mount. After drilling two small holes spaced for the tabs in the Powerade lid I roughed it's surface along the top outer edge with an emory board. I also carefully roughed the top edge of the heatshrink on the capacitor. The purpose of this is to eliminate smooth surfaces for better epoxy adhesion. With the lids attached and the epoxy cured, I put a small piece of red heat shrink on the positive lead of each capacitor as an identifier. As a final prep for installation I took one of the original caps and used it as a guide to bend the leads of the new caps into ideal position and spacing for installation.


A comparison of the original cap on the left, and it's replacement, right.
GEMO369.jpg


Powerade cap atop the capacitor to act as a spacer. 22 AWG wire wound around the solder tab serving as a through hole lead.
VkmOR3O.jpg


Two capacitors almost ready for installation. The leads soldered on, flattened, and formed to the proper spacing.
LPIgEoF.jpg


Edit Additional thoughts on the mounting of the filter capacitors.

The mounting of diode boards is solely reliant on the filter capacitor brackets. This is important because the position of those boards have only about a 1/8" sweet spot for them to avoid shorting grounding pins and components to the chassis, potentially disastrous. Therefore it is critical that the main filter capacitors be FIRMLY clamped in the brackets.

The Kemet capacitors I used, and pretty much any other modern 40mm diameter replacements that you may want to use, will undoubtedly be about 1-2mm smaller in diameter to the originals. In order to ensure a snug fit I would strongly recommend wrapping the caps with a material based tape a few times around where the bracket tightens around, hiding it from showing. A tape such as this will serve to grip rather than slide. Gaffers and athletic tape are some common examples.

This image shows the proper mounting height. The boards' pins are clear of the chassis by about 1/8" . I marked the filter capacitor polarities on the chassis prior to installation since there are none silkscreened anywhere on the boards.
HlysFNu.jpg
 
Last edited:
Interesting mod.

In my best southern draw :
It ain't purdy, but I rectum it'll get the job done. Duct tape, bailing wire, and a rock fer a hammer 'bout all you need to fix all you need...ceptin your seeg-neefa-gunt utter.
 
The foam used in these 70's Kenwood's is always starting to dry rot. I make a point to remove it because if you don't it begins to disintegrate into particles that crumble into switches and pots and start resulting in issues. I've been using adhesive backed craft foam to replace the old stuff. It comes in some nice 9" x 6" inch rectangles in various colors which I trim to size.

Old foam from the meter assembly mostly removed
tKFoaex.jpg



The meter assembly cleaned of all the old foam, old adhesive, some crumbling debris still remains.
qHVuKlH.jpg


Preparing a piece of craft foam to be trimmed.
J8CCwQR.jpg


The strip of craft foam applied in the same manner as the original.
Jj0ns59.jpg


The light assembly has a foam strip on it as well so the same approach was taken.
0ZCi75j.jpg


A nice clean piece of foam is applied. No more crumbling foam.
KJ3hWXW.jpg


 
Last edited:
Hey John:
Good idea. I used some very thin weather stripping in a bind or similar to what you are using.

John M
 
Addressing the KA-9100 Service Bulletin

There is a small 3-pin header on the top control board which is subject to strain from the connected cable and results in intermittents. Kenwood issued a service bulletin to deal with the issue which directs the tech to remove the header and connector and solder the wiring directly to the board. My suggestion instead would be to provide strain relief to the header with some adhesive or hot glue strengthening its bond to the board followed by a thorough redo of the three associated solder joints. The larger 6-pin headers don't seem to be as effected.

yweoE91.jpg


MTiV4jF.jpg


i1oJeHS.jpg
 
Last edited:
There were several wire wrapped connections which had broken. This occurred both as result of disassembly, and from continual board movement over the course of component installation. These wires were replaced with same gauge solid core wire. Each was wrapped onto the posts, mimicking the originals, and then soldered in place.

Broken wire wraps are nearly inevitable to avoid when changing the number of components I've done on these boards. I attempted to make them appear as original as possible in my repair.
m7yA8qg.jpg


Once the foam replacement, as documented, was completed on the two assemblies the VU meter wiring was reattached to its board. This board has a shield which was glued to it along the right edge (facing front) which had fallen off. Once remounted, I used a bit a hot glue to reattach the shield as it was meant to be.

The control boards, small switch boards, diode boards, were all cleaned of flux and reassembled into the chassis.

The amp was powered on with the dim bulb tester. After testing good, it was powered on in earnest checking for functionality and for a clean signal among all controls. In listening, there was a noticeable improvement in audio quality, primarily in the authority of low frequency reproduction.

Problems with the volume pot start to exhibit themselves. With the pot turned turned to minimum the volume would jump wildy up in db and sometimes drop the right channel. The spray and pray method with F5 wasn't cutting it. Since only portions of the control can be reached with the board mounted it needs to be removed to expose the bottom of the control. After a flooding of CRC in the forward volume portion of the control some small debris was freed from the control which seemed to be the source of my issues. This was followed up with applications of F5 and G100L and the requisite 20-30 turns of the controls.

Entry points for cleaner are exposed from the bottom of the combination balance/volume potentiometer once removed from it's PCB.
wXPCS8c.jpg



Phase one of the restoration is completed.

Up next, phase two of the restoration begins with the power supply.
 
Last edited:
No matter how many times people have said it, I am just stunned at how hard the KA-9100/9150 is to wrestle around and work on and amazed at those of you who've done a great job of it already. I'm... wrestling my way to getting mine all back together after pulling everything out, recapping and upgrading. Man!

IMG_5465_edited.jpg
 
I see you have one of the rare 9100s with four transformers, Doug. :biggrin:

Hah! Nice catch! I'd bought a spare set just in case these were toast. I'm 80% done with the restoration, so it was time to remove the tranny covers, sand blast them, and repaint before bolting them back down in the 9150. I pulled out the spare transformers so I could do all four covers at one time.

The only advice I can give you is don't walk away from it! It's not like each board has four screws and they neatly bolt to the chassis, or that the components are easily accessible from the top and bottom side of the boards when the boards are mounted, or that all wiring has molex connectors. I'd love to see the instruction set given to the assemblers for exactly how they were instructed to pack those diode boards in there. :eek: Beyond that, it's cake. :)

No chance of turning back now. I bought this unit last year for myself, knowing how much acclaim the KA-9100 gets and how cool the Euro version looks. I was considering whether or not to dedicate this coming weekend to finishing it up when someone in BT posted "WTB KA-9100... recapped". That and a COBRA health insurance payment coming up made the decision easy.

If you all see any posts from me in the next few days saying "help! where the heck does the pink wire go?" please be kind to me. This isn't my first restoration, not even my first Kenwood, but it is certainly the trickiest. I've got a pair of KR-6160's and a KT-140X's opened up where I stopped and said "whoa, THAT is gonna take serious attention and time." The 9100/9150 is denser in design and WAY harder to physically wrestle around the bench.
 
If you all see any posts from me in the next few days saying "help! where the heck does the pink wire go?" please be kind to me. This isn't my first restoration, not even my first Kenwood, but it is certainly the trickiest. I've got a pair of KR-6160's and a KT-140X's opened up where I stopped and said "whoa, THAT is gonna take serious attention and time." The 9100/9150 is denser in design and WAY harder to physically wrestle around the bench.

dlucy---->> Just start another thread on your restore. Gets confusing to others that follow along.
Thanks
John M
 
No matter how many times people have said it, I am just stunned at how hard the KA-9100/9150 is to wrestle around and work on and amazed at those of you who've done a great job of it already. I'm... wrestling my way to getting mine all back together after pulling everything out, recapping and upgrading. Man!

View attachment 1008812
It wasn't as hard as everyone said. One board at a time: power supply first, then amp, then preamp, then main cap board. I have a hakko 808 so maybe that helped. Made a mistake and was able to find it (broken wire). I didn't have to take the whole chassis apart, like in your pic Good luck.
 
The KA-9100 power supply in stock form. (X00-1870-11) This is the K version of the power supply for North America. Foreign or multi-voltage supplies may differ.
0ojSU5Y.jpg


Changes made to the KA-9100 power supply board (X00-1870-11) have been completed.

After the changes were made a DBT power on was performed. After a glow and dim on the DBT a full power up was performed. The amp took a full 20 seconds or more to come out of protection and the VU meters were reading 10W. Holy Moly, what's going on here? Rather than panic I applied the KISS method. I know it worked fine before working on this board so it has to be something wrong with the supply. After a thorough recheck of every component changed I was confident nothing was replaced incorrectly. I turned my attention to the bottom edge of the board to see if there were any broken wires and again found nothing wrong. As I was moving the board looking things over I spotted pin 25, the -B supply, had a cracked solder joint. Bingo. I re-flowed all the pin joints just as precaution and then powered up again. After 3 or 4 seconds, PING! relay engaged.

The KA-9100 power supply with new components installed.
PyYB6fv.jpg


Rk11 1K 2W was replaced with a 3W metal film and given strain relief. The two stock 470ohm at Rk2 and Rk4 were replaced with 1W KOA Speer SPR Carbon Film replacements ( I believe the originals may be flameproof type so I went with the SPR which have flameproof coating ) and their leads given strain relief. All the other resistors were replaced with the 1% .5W metal film. I made point to raise the 820 ohm Rk1 and Rk3 off the board a bit for better heat shedding. All the W06B diodes at Dk1-5 were replaced with 1N4007G rectifiers, raised a little off the board. The YZ-140 at DZk1 was swapped for a 1N5244B Zener.

There are ten electrolytics here.

Ck11 4.7uf 35v (25v actual) > MKS2 4.7uf 50v
Ck14 47uf 16v 8mm > 47uf 25V RFS 8mm
Ck8 47uf 35v 10mm > 47uf 35V RFS 10mm
Ck6 47uf 50 10mm > 47uf 50V KZ 10mm
Ck10 10uf 25v > 10uf 50v FC
Ck5 100uf 50v 12.5mm >100uf 63V PM 12.5mm
Ck12/13 220uf 35v 12.5mm > 220uf 50V FC 12.5mm
Ck3/4 470uf 50v 16x35mm > 560uf 63V HE 16x30mm

The four transistors were replaced as follows

2SC1983 > MJE15032G
2SA755 > MJE15033G
2SA673 > KSA1220A
2SC945 > KSC945CGTA

There are some points of interest to working on this power supply. It should be noted that Ck16, a 47 picofarad ceramic cap, is on the rear of the board. It is soldered across the collector and base of Qk4 2SA673. If you replace Qk4 you'll need to remember to put this cap back in place. My board had glue over the solder joints so you may need to remove that first.

rlhCpoh.jpg


Ck16 47pf

FoVgzI3.jpg


Also of note, the regulators at Qk1 and Qk2 are mounted to heatsinks which mount to the bottom of the PCB. During removal it was obvious that all of these screws were very loose and might have easily come unscrewed given time. Whether you replace these transistors or not I would suggest tightening them up and adding some thread lock to prevent any of them from falling out. These same transistors also did not have any thermal compound on them. I would suggest adding a thin layer for better heat dissipation.

Thread lock is added to the transistor mounting screw. Some should also be applied to the board mounting screw.
QbvwSkc.jpg


A thin layer of thermal compound is applied between the regulator and heatsink prior to mounting.

7mNO69n.jpg
 
Last edited by a moderator:
Good work John, and good catch on the cracked solder joint. I was going to mention the possibility, but thought 'nahhh'..... I had the same thing happen on mine and only noticed it when I saw and heard the arcing. I forget which pin, but I think #1. I also reflowed all of them.

Interesting note on Ck16 - I didn't remember seeing that. Looked at the photo I took and found it mounted on the top side. Must have been a board change at some point or due to mine having switchable voltage, who knows.

P1011024.jpg
 
Back
Top Bottom