One of the amplifiers I own is a Kenwood KA-7100 stereo amplifier. It is a fairly old amplifier (late 70’s), with a modest 60WPC into 8ohm specification. I was a fan of the KA-XXXX amplifier series and its siblings from the moment I first heard its smaller brother (KA-601). While that KA-601 was in fairly bad shape, I was amazed at the time by how much better it sounded than my (back then) modern Denon AVR HT receiver. I’ve since had an opportunity to listen to quite a few amplifiers from this series including the KA-7300, KA-8100, KA-9100, and others. Therefore, when I’ve had the chance (more than a decade ago) to get my hands on a KA-7100 I grabbed it right away. It wasn’t in bad shape, but it has seen better days, no doubt. As I learned a bit more about electronics, that amplifier became one of my first projects.
I have recently had the opportunity to revisit this amplifier by doing another small modification, which gave me an opportunity to write something about it, and add some measurement results while at it.
Back when I just started modding this amplifier, I’ve noticed someone had worked on it already. However, they did a very bad job at it. In fact, it was so bad, that one of the Zener diodes in the power supply section was replaced with a simple diode in reverse bias. This means that this specific rail was both too high, and too noisy. Other than fixing the problems it had, I did a few extra things to it:
- Replace spring loaded speaker connectors with Banana jacks
- Replace all electrolytic capacitors with modern equivalents
- Replace rectifier diodes with low voltage drop variant
- Replace RCA connectors (which were slightly corroded) with gold platted connectors
- Replace output relay with new Omron relay (old unit had developed intermittent contact issue over the years)
- Replace volume and balance potentiometers with Alps RK27 type
- Remove the “TAPE” selector switch and bypass it on the board
- Add shielded cable for line level signal from input selector board to front panel board, and power amp board
- Add regulator “boards” for low current supply rails
- Take apart and thoroughly clean each and every switch to extend the contact life time
- Adjust bias to target value (it has drifted noticeably over the years)
The regulators I’ve used are a simple variant of the regulator described in this post. These regulators started out as mods for vintage amplifiers like this KA-7100, and I’ve only later developed them further into a more complicated circuit on its own PCB. These regulators have poor thermal stability, but offer very low noise with good line regulation, which is a good fit for audio circuits.
One additional mod that I did was removing the second output and leave just a single output option. This meant there was no more need for the A/B/A+B selector switch which I’ve taken out of the circuit. Some years have passed since these mods were implemented. I’ve since replaced this amplifier with the B22 amplifier I’ve built. However, recently I’ve decided that I would like to add the second output option back to the amplifier and get this amplifier back to use in a different system, which meant revisiting it.
Other than creating a new small aluminium back-plate with the 4 pairs of Banana jacks mounted to it, I’ve taken the opportunity to clean the A/B switch thoroughly, and replace the output wires inside the amplifier with a higher diameter variant. Than, finally, I’ve used this opportunity to place the KA-7100 on the bench to measure its distortion and max output power before clipping, as I didn’t have the gear to measure this years ago when I’ve modded it originally.
To measure it I’ve connected to my normal setup of EMU 0404 USB sound-card + measurement pre-amplifier. Both channels were loaded with 8ohm dummy loads, and I wanted to measure 2 numbers. The first, is the actual distortion at the rated output power, the second is the maximum output power before onset of clipping
The specification of this amplifier states a distortion of 0.02% at 60W output power into 8ohm. The following two figures show the actual output spectrum and distortion at 60W (22Vrms), and 72W (24Vrms) which is the maximum swing it was able to achieve before clipping drove distortion upward very quickly.
The results are actually not bad. At 60W output power, the distortion is just 0.0028%, which is almost X8 lower than the number found in the spec. Driving this to the point of clipping we can see that the maximum output power with low distortion is 72W, which is a nice 20% overhead compared with the spec.
I’m quite happy with these figures out of this old horse, it will serve well in the new setup it is going to be a part of.