The KA-2000 restoration

The main amp output is indeed cap coupled - 1000uF/25V - you can see them in photos 6 or 7 in post #1
 
You may want to bump these up to 2000uF, but I'm sure hopjohn has a plan.

Increasing the capacitance value in the output-coupling caps will dramatically increase bass emphasis. I very much so like increasing capacitance in a rebuild, but the output caps are the one place with a negative outcome.
 
From what I've read, 2000uF is the sweet spot for a coupling cap. But I've been wrong before, so I'm interested to hear what hopjohn weighs in with, he always has carefully considered the options.
 
You may want to bump these up to 2000uF, but I'm sure hopjohn has a plan.

Increasing the capacitance value in the output-coupling caps will dramatically increase bass emphasis. I very much so like increasing capacitance in a rebuild, but the output caps are the one place with a negative outcome.

From what I've read, 2000uF is the sweet spot for a coupling cap. But I've been wrong before, so I'm interested to hear what hopjohn weighs in with, he always has carefully considered the options.

Gentlemen, I'll likely be increasing them to 2200uf 35V or 50V which seems a safe value. That gives me a -3db point at about 9hz. That should be low enough to have some buffer for reaching down to 20Hz without a lot of rolloff. I certainly don't want to go crazy here with the increased capacitance which could have a negative impact, but the 1000uf does seem a little small in value and may have been decided upon due to physical size at the time which isn't an issue nowadays.
 
More on 2SC458 replacement in the Main Amp 8215

9tULFjd.jpg


The AF amps, Qe1 and Qe8 of the 8215 amp board, are the infamously noisy 2SC458 and should be replaced. The schematic doesn't reflect it, but each of these transistors has a 100pf ceramic cap across Base-Collector. The 100pf cap shares the through holes on the board with the base and collector legs.

The outhouse styled package, 2SC458 manufactured by Hitachi has an ECB pinout and is read left to right with the smaller side and slope towards you. The ID is on the reverse as shown in the photo above.
 
Last edited:
Main Amp Board 8215 continued

Completed amp board
5JfBkas.jpg


Center Voltage and Bias Adjustment

Once the 8215 Main Amp board of the KA-2000 is rebuilt six new trimmers will have been installed. Prior to beginning any restoration work I made all the necessary amplifier adjustments in stock form to provide a starting point for the work ahead. While working on the amp board each original trimmer setting was measured and the values duplicated for each of its replacements. Even still, the first power up should be done on a dim bulb tester (DBT) with a bulb about twice the value of the W/per channel. Here we have a 13W/Ch amp into 8 ohms so I used a 28W bulb.

Once powered up with the DBT you can check the center voltage for each channel. First, use a DMM to measure DC voltage across the collector (just place your red probe to one of the transistor mounting screws for the collector side) and chassis ground of output transistor Q2. Record the value and then repeat for Q3. Using the DBT will limit the current some so your voltages will be a little under 42V in the 39-40V range. Likewise measure and record the collector voltage of Q1 and Q2. The reading for each of these should be a little under 21V and about half that of the corresponding channel in the 19-20V range. Verify that Q1 is close to half the value of Q2, and that Q4 is close to half the value of Q3. If adjustment is required use VR1 for setting Q1 (left channel), and VR2 for setting Q4 (right channel).

The output transistors as seen top side. face forward, from above. The "C" is for collector.
q5l0Kob.png


Still using the DBT, we set a preliminary and safe bias setting for each channel. We do this by measuring the DC voltage at the emitter of Q1 and ground for the left channel. Repeat by measuring the emitter of Q4 and ground of the right channel. At full power the bias should be 14.1mV so while using the DBT set it to about 70% of that value to be on the safe side thus around 9.9mV. Trimmer VR5 adjusts the left channel and VR6 adjusts the right.

Now that the preliminary settings have been made it should be safe to turn on under full power. Allow 20 minutes for the amplifier to get up to full operating temperature. Using the procedures previously described, set the center voltage. Again, the voltage of Q1 should be set to half the voltage of Q2 using VR1 and the voltage of Q4 should be set to half the voltage of Q3 using VR2.

Once the center voltage is set you can move on to the bias adjustment. With the amp at full operating temperature set the bias to 14.1mV for each channel. Red probe to the emitter of Q1, black probe to ground adjust trimmer VR5 for the left channel. Red probe to the emitter of Q4, black probe to ground adjust VR6 for the right channel.

Since we have yet to work on the power supply (filter caps, rectifier etc.) section at this point in the restoration it would be a good idea to revisit the adjustments when that has been completed.

Note for the geeks: The bias current for the KA-2000 is 30mA. This being a quasi complimentary design we measure across only one emitter resistor (R11 L, R12 R) to find the bias. The proper voltage can be calculated by the following: 30mA x 0.47uf = 14.1mV. Quasi complimentary designs were used in the day because the performance of complementary NPN and PNP transistors wasn't what it is today and the reason you rarely see it anymore.
 
Last edited:
My 2000 tone board had r1 marked as r7. Had wrong resistor there when I replaced them all thanks to that. Also has 2sc350 to-1 in tone board.

Good call on headphone jack cleaning.
 
My 2000 tone board had r1 marked as r7. Had wrong resistor there when I replaced them all thanks to that. Also has 2sc350 to-1 in tone board.
You must be thinking R2, it's mislabeled as R7 on the tone board. Your unit must be a little older than mine. I believe the 2SC350 T-01 type were used early on in the production run.

Good call on headphone jack cleaning.
Thanks, I could see it being a potential issue.
 
Power Supply Part 1

The two stock rectifiers are soldered in parallel with .022uf caps on a terminal strip mounted to the bottom of the chassis.
8EHzJTq.jpg


New 1N4007 rectifiers and .022uf caps installed
766iKEX.jpg


Rectifiers

D1/D2 > DS1N rectifiers (FR2-? actual) > 1N4007
C12/C13 (not in parts list) > .022uf 630V axials > Wima MKS4 .022 630V
 
Last edited:
Power Supply Part 2

The total original filtering capacitance is only modestly increased from 3420uf to 5020uf. I intentionally did not want to get carried away here. The large filter is 35mm in diameter and the smaller 22mm. These were replaced with like sizes to make use of the original clamps without any needed modification. The 220uf axial cap at the bottom of the chassis was replaced with a 220uf radial. The original resistors in the circuit were replaced due to the point to point nature of their installation. Metal film resistors were used in place of the original carbon composition types of the same values.and wattages.

The new 1500uf 50V cap was given ring terminals to help accommodate all the required connections, Similarly, the 3300uf 63V cap was given crimp on pin terminals.

Original Filter Capacitors
wIKz3aR.jpg


Original Filter Capacitor Wiring
yDW63Cl.jpg


New Filter Capacitors
T9KKvkZ.jpg


Ring terminals attached to C10
fAX5Bjg.jpg


New Filter Capacitors Wiring
bomZPFW.jpg


Filters Capacitors
C9 220uf 25V > 220uf 35V Nichicon HE (Terminal strip to - of C10)
C10 1000uf 30V 22x40mm > 1500uf 50V Nichicon KG 22x30mm (Ring terminals added)
C11 2200uf 50V 35x50mm > 3300uf 63V Nichicon KG 35x30mm (Pin terminals added)

The following replacement resistors were all metal film type:
R17 220 Ohm 0.5W (Terminal strip to + of C10)
R18 4.7K 0.5W (Across +/- of C10)
R19 1K 2W (+ of C9 to + of C10)
R20 47 Ohm 0.5W (Bottom, center terminal of power switch to + of C10)
 
Last edited:
Output Coupling Capacitors

The original output capacitors were a little smaller than they needed to be in my opinion, possibly due to cost or physical size restraints when built. An increase to 2200uf should provide better low end response with no compromise.

Original Output Coupling Capacitors
rYrsnA9.jpg


New Output Coupling Capacitors
QAttfQ9.jpg


Output Coupling Capacitors
C5/C6 1000uf 25V axial > output coupling > audio grade > 2200uf 50V Nichicon KW
 
Last edited:
Is this yours, or a customers? I can't get over how pristine the interior of that amp is!
This is for my personal collection. I was fortunate to have found a well cared for example. It likely spent most of its life inside a box as there was little to do to clean it but dust it off a bit.

As usual, nice work!
I have to admit It's not my cleanest work. I haven't had the time I'd like to get it completed, but I appreciate the kind words nonetheless.
 
Output Coupling Capacitors

The original output capacitors were a little smaller than they needed to be in my opinion, possibly due to cost or physical size restraints when built. An increase to 2200uf should provide better low end response with no compromise.

C5/C6 1000uf 25V axial > output coupling > audio grade > 2200uf 50V Nichicon KW

Can you estimate the bass emphasis? The first time I bumped up output caps, I didn't realize how much mine (hk330b, I think) would produce. I kept swapping speakers, thinking that was the issue. Then turning down the bass.
 
Can you estimate the bass emphasis?
The low frequency response improvement should be subtle, and maybe not even audible depending on the speakers used. I'm not familiar with the HK330b, but depending on the design of the filtering network it may not be a good idea to be changing values.in that model.
 
Last edited:
Miscellaneous

Foam/ Light blocker
There is a piece of foam near the lamp used to prevent light leaking through the faceplate at the switches. This foam deteriorates and causes a dusty mess that can get inside switches. As preventative maintenance I remove and replace this stuff.

Old, rotting foam.
DVTsI6S.jpg


New Lamp Foam
XX92cH3.jpg


Pilot Lamp (Power Indication)
The schemaic claims a 7.5V secondary for the lamp. The stock bulb is a Stanley 8V 0.15A with bayonette base. For a replacement I'd try a widely available #51 7.5V 0.22A CEC industries with a BA9S base. If the #51 proved to burn out too quickly I'd try the #1408 10V 0.13A CEC Industries with BA9S base. These are both available at memotronics.com

Safety Cap
It's a good idea to replace the safety cap with a modern UL approved X1Y2 type. The safety cap is soldered across the fuse holder and one side of the convenience outlet at the rear of the KA-2000. If my Trio brethren are on 220-240V you may want to find a safety cap with a higher voltage rating than the one I've used for my U.S. based 120V.

Original Safety Cap
0iQ3lrq.jpg


New Safety Cap
XOuFSHg.jpg


Safety Cap
C14 .01uf 500VAC > .01uf 250V Murata X1Y2 Rated Mouser # 81-DE2F3KY103MN3AU2F
 
Last edited:
Summary

Restoring this KA-2000 has been a fun project. All of the restoration goals were met as listed in post #10. From a service perspective, the boards are relatively easy to access since there is both a bottom and top cover. The biggest difficulty, if you want to call it that, is desoldering the bent over component leads that run over to other components on the same pad that aren't being removed. In total there were 38 capacitors to be replaced.

Upgrades were made to a number of the electrolytic capacitors with polyester film caps and the potentially noisy 2SC458 transistors were given modern replacements on all three boards: tone amp, preamp (phono) and the main amp. The power supply was upgraded some with new modern rectifiers and the filtering was modestly upped from 3420uf to 5020uf to provide a little more reserve.

The output coupling capacitors were increased in capacitance from 1000uf to 2200uf to lower the -3db frequency point to around 9hz. The stock -3db point looked aimed at 20hz. I should point out that I'm not sure how much improvement the 2200uf caps will provide with the input caps of the other boards remaining at stock values. This is perhaps something to explore in the future. In any case, with the mods completed, the amp was fed a sine wave while being observed on the scope. The amplitude of the sine wave remained consistent right up until 25hz and only then began to dip off slightly as the frequency of the AG was lowered. Listening tests confirm the KA-2000 has no problems producing clean bass.

My biggest complaint with the KA-2000 would be that it uses germanium transistors in the main amp that aren't directly replaceable with silicon. I feel these transistors are at least partly to blame for the slight channel imbalance that was exhibited.

All in all this little amp will prove to be a great piece for anyone looking to save on space with its small foot print. A definite upgrade for any computer arrangement. It is also well suited placed on an office desk or in a small room setup. I have much enjoyed listening to it through my favorite cans, a set of Status Audio CB-1. You can be sure it'll make a for a great headphone amp. It's simply a fun, classic amp that is also an interesting landmark in Kenwood history.

rXHGZ66n_o.jpg
 
Last edited:
Another top-notch restore and write-up! Thank you from all of us!

My biggest complaint with the KA-2000 would be that it uses germanium transistors in the main amp that aren't directly replaceable with silicon. I feel these transistors are at least partly to blame for the slight channel imbalance that was exhibited.

It would be really interesting to hear similar-designed units, A/B, to prize out any sonic differences between Germanium and Silicon. They have different operating characteristic curves, so there must be differences. Ive just never had the chance to compare such units. Im terribly curious on this. Any thoughts to share?
 
Ive just never had the chance to compare such units. Im terribly curious on this. Any thoughts to share?
With regard to germanium vs silicon, I've not heard or read much about the differences in sound reproduction for home audio. However, as with so many guitar enthusiasts who are chasing a certain tone, I've read a number of things about how germanium transistors provide a smoother distortion for them. That doesn't translate well to our hi-fi world, but it at least gives you the impression that there are real differences in how they sound. In my experience, all the cap coupled amps I own have a sound flavor closer to that of tube designs than those that later used direct coupled complementary outputs. Which seems to make sense as these same amps are only a few years removed from when tube amps were still in full production.

Edit: Found this video that demonstrates the differences from a guitarists point of view.
 
Last edited:
Back
Top Bottom