img 9697

Legendary Krell KSA 100 Class A Power Amplifier Restoration

Welcome back folks, it’s Monday and that means it’s both my day off and article time, darn it! Anyway, this time it’s another amplifier restoration. We’re heading to a similar time in the 1980s, but with a different manufacturer, Krell, and to the OG myth, the legend, the Krell KSA 100 class A power amplifier!

History

The Krell KSA-100 is one of the original amplifiers that put this legendary company on the map. It is a classic Dan D’Agostino design and a true class A beast, verified by its power consumption, as you’ll see. These amplifiers appeared in around 1981 and were produced until the very late 1980s.

This model utilises active cooling in the form of fans rather than the passive cooling solutions used later. One could make an argument that this worked better and it certainly allowed the amplifiers to be lighter and more compact. They only worked better until the fans died of course. Then there were real problems.

Krell KSA 100
Yes, she’s old and a bit battle-weary, but she’s a Krell KSA 100 and that means she resides in the upper echelons of legendary amplifiers, regardless of how she looks! This KSA 100 is a late one, from around 1988 I think, based on date codes I found.

The KSA 100 came shortly after the also OG Krell KSA 50 and was followed by the beefier KSA 200. Many other models followed of course and you’ve seen me work on many of them here. These models include my old KSA-150, the KSA-100S and the KSA-200S about which I’ve yet to write a long overdue article, sorry Paul!

img 0549 1
The very lovely Krell KSA-100S, an amplifier and series I am very fond of.

I’ve had Krell factory support, mainly in the form of service data, since I started Liquid Audio. It’s fair to say that my articles have helped raise awareness and popularity of these wonderful Krell legacy models and it’s been a pleasure to contribute to the Krell heritage. Krell now seems to prefer to fix legacy equipment in-house, but this was not always the case.

When I most recently (2023) asked for service data for this particular KSA 100, Krell didn’t respond, the first time that’s happened in 15 years. I really appreciated it when I could email Patrick Bresnahan or Ray Mutchler and someone would get back to me within a day or two. Perhaps those days are gone, but I remain optimistic.

A colleague kindly supplied me with the complete service data for this one and I had some of it anyway. Perhaps this recent experience was just a glitch, but I’ve been saving Krell equipment like this and undoing bad repair work on it for a long time. I hope Krell doesn’t forget those of us who’ve worked hard over the years to help keep their legacy equipment alive.

By the way, if you enjoy content like this and don’t want to miss it when I post new articles, subscribe!

Krell KSA 100 Specifications

These are KSA 100 Mk2 specs, adapted from the Audio Database

TypeStereo power amplifier
Rated output100W + 100W (8 Ω, both channels operating, 20 Hz to 20 kHz)
200W + 200W (4 Ω)
400W + 400W (2 Ω)
800W + 800W (1 Ω)
Input Sensitivity / Impedance1V/22kΩ
Total harmonic distortion factor0.05% or Less (20 Hz to 20 kHz, 1W to Rated Output)
Frequency characteristic
(1W ~ Rated Output)
0.5 Hz to 100 kHz -3dB
20 Hz to 20 kHz -0.5 dB
Signal-to-noise ratio (IHF-A)110dB
Rated load impedance1 Ω or more
Damping factor120(1kHz)
Slew rate80 v / μ sec
Power consumption610W (maximum idling) Yes, seriously!
External dimensionsWidth 483x Height 226x Depth 612 mm
WeightApproximately 35 kg

Issues

The KSA 100 is a proper class A amplifier and that means it generates a lot of heat. Class A amplifiers sound the best, but cannot be kept in cabinets or under low shelves, though they often are. The fans in these models are known to fail after spinning millions of times over decades. These amplifiers need regular maintenance and as all my regulars will know, herein lies the crux of the issue.

Krell KSA 100

These amplifiers and others like them usually either don’t receive that maintenance for various reasons, or receive poor quality maintenance and fail. Big class A amplifiers often fail because of neglect or poor service/repair workmanship. This amplifier exhibited a bit of both.

This old girl was in average condition inside and out, suffering from the effects of corrosion, heat and previous poor work over extended use. Keep in mind that this particular unit, a late ’80s serial number, is 40 years old. Four decades of pulling nearly 3 Amps from the wall is going to have some negative effects.

She had let go in a big way at some point in her past though, as evidenced by the new output devices and emitter resistors, various replaced small-signal transistors, and so on. When you combine the poor work from previous repairs, the broken threaded rod/fan assembly, dry thermal interfaces, unsoldered components, cheap emitter resistors and dead caps, you have the perfect recipe for a bad-sounding and poor-running KSA 100. That’s exactly what this was.

The issues I found included:

  • Need for typical age-related maintenance
  • Broken threaded rod securing fan/heatsink assembly
  • Amplifier and driver modules need overhaul
  • Output device thermal interfaces are highly problematic
  • Filter capacitors are tired
  • Many fasteners are wrong, worn or in the wrong locations
  • One driver transistor was not even soldered in place!
  • High DC offset, etc, etc

My brief was to repair and restore her, with a maximum budget that I of course adhered to. I was able to address and resolve all these issues within the owner’s budget.

Restoration

This is a long and involved job. I’ll need to include many photos, but I’ll try to keep things reasonable as I don’t want to bore everyone to death. Also, note that the budgetary discussion with her owner meant that I could not attend to everything I might like to do here. I’ve deliberately left some things where it was safe and sensible to do so.

The emitter resistors for example have all been replaced at some point and measured fine. They are not the premium Dale parts I fit when doing this work, but we typically cannot do everything unless the owner gives me a blank cheque! That’s never happened, but there is always hope 🙂

Let’s go!

Krell KSA 100
Lid off, I’m afraid to see what lurks within! Note that the Krell KSA 100 is a true dual mono power amplifier, with one transformer, two diode bridges and two big filter capacitors, PER CHANNEL. The only shared element is the mains wiring.
Krell KSA 100
If you know what you’re looking at you’ll see a lot of dirt, dust and replacement resistors, plus the usual age-related corrosion, wear and tear and some replaced wiring.
Krell KSA 100
You know you have a serious amplifier when you see two 25A diode bridges – PER CHANNEL! They are cleverly thermally bonded to the ground bus between these two filter capacitors and there’s some associated heavy-duty wiring. I’ll replace all these old fasteners because nobody else will!
img 3901
I’ll also replace these worn-out non-original philips-head fasteners, again, because nobody else will bother!
Krell KSA 100
So that’s four 25A diode bridges in total, needed when you have a stiff power supply design where power output doubles as load impedance halves. I’ve worked on amplifiers that claim big power and have just a couple of puny 10A bridges, which is a joke.
Krell KSA 100
Stripping her down gives me access to various bits and pieces and allows for a deeper clean, coming up.
Krell KSA 100
This is one of two DC protection boards that will be overhauled. These capacitors will go.
Krell KSA 100
Trying a new iPhone filter that removes the background, that’s not too bad for an AI-type process! Dirty Krell KSA 100…
Krell KSA 100
And now sparkly clean after applying my deep-cleaning regimen. What an improvement.
Krell KSA 100
Pulling all this apart when stuff is broken and fasteners are incorrect and worn out is not the most fun of jobs. It has to be done though. This is an amplifier module, here we are looking at a driver board. Many fasteners have specific lengths and were installed in the wrong places by the previous tech. This is ‘normal’ but annoying as all hell. Every. Single. Time.
Krell KSA 100
Here we see the fan bracket and broken threaded rod. The previous repairer probably broke it, and then just left it, because who cares, right?!
Krell KSA 100
Disappointing, but honestly, nothing surprises me these days.
Krell KSA 100
Here, I’ve overhauled the driver and DC protection boards for this channel. Note the new, premium light temperature, low-impedance, high-ripple current rated capacitors, and new thermal pads for these driver devices.
Krell KSA 100
I can’t recall if this is the same or the other channel board, but you get the idea and my work was applied evenly.
Krell KSA 100
All output devices and heatsinks were in this state. Devices were mounted on silpads that shouldn’t have any thermal paste. The thermal paste that was used had dried up decades ago anyway, very bad for a class A amplifier running such high quiescent current. I’ll clean everything and remount all transistors on new high-performance silpads.
Krell KSA 100
You’ll not from the date codes on these devices that they are replacements, but they all measured well, so I re-use them on lovely new Berquist thermal pads. I assume the bad work I found here was done in or around 1999/2000, or not long after. What an improvement.
Krell KSA 100
My secret weapon in this build: uprated RIFA/KEMET filter capacitors. These mil-spec industrial-grade capacitors are the very best you can buy, with almost indefinite shelf and service life and the highest rating for temperature and ripple current available. Superb. Beefy little suckers too, it’s like holding a can of beer.
Krell KSA 100
The scale and complexity of working on amplifiers like this means repairers able to take it on need the time and space to do it. This Krell KSA 100 owner completely understood this.
Krell KSA 100
Modules are coming together nicely. These output devices will have a much better time of things now, as the heat they dissipate will be conducted into the heatsink and then the surrounding air. Low thermal impedance is critically important in a design like this.
Krell KSA 100
Look what I found in the second channel though. This was so incredible to me that I couldn’t believe it at first. This driver device was just sitting on the heatsink unsoldered into place, not connected to the circuit. Because Krell runs parts like this in parallel, the other one is doing the work of two. Naturally, I rectified this, but imagine not checking your work carefully enough to see you’d made this mistake, or maybe not caring enough to bother.
Krell KSA 100
Resolved, with new thermal pads used as per the other channel. You’ll note some correct but non-batch-matched parts here. For this brief and budget though, this is acceptable.
Krell KSA 100
Reassembly of the module without the damaged fasteners/threaded rod is now possible. Honestly, these modules are a pain to work on, but Krell improved their designs over time as one would expect. Note the non-original power resistors.
Krell KSA 100
Here we see a completed module with associated DC protection ready to be reinstalled in the chassis.
Krell KSA 100
You know what they say, measure three times… We don’t want any of this DC power circuitry shorting against the lid. Perfect fit.
Krell KSA 100
These RIFA capacitors will transform the sound of this amplifier.
Krell KSA 100
Beautiful! You may not be able to see it but I am using new fasteners, star washers, etc.
Krell KSA 100
It just feels good doing things correctly, doesn’t it?
Krell KSA 100
I remade these speaker terminal connectors and re-heat shrink-wrapped them. For those wondering, this wiring is not original.
Krell KSA 100
I also re-made these speaker terminal networks using premium new capacitors – you’ll note the broken leads on these.
Krell KSA 100
Perfect, and made with parts from stock. Note the cap-to-resistor soldering, on the left. It helps to have deep stocks of parts like these, just in case.
Krell KSA 100
Parts are now nicely heatshrinked, job is done.
Krell KSA 100
Now THAT’s what I’m talking about. I’ll bet this amplifier sings after this (spoiler: it does!).
Krell KSA 100
Time now to tackle this broken threaded rod. I purchased some new rod of the correct original spec and have now screwed it into the already installed amplifier module. This sounds so easy when you read it here doesn’t it..? Screwing the rod in was painful and screwing the nut onto the rod was more painful. You’re wondering “Why not cut the rod first, then screw on the rod..?” Well, if you can’t screw on the nut, you’ll have to extract the rod and that would have been very difficult in this case. Hey, it only took about 10 minutes to screw on the nut… 🙂
Krell KSA 100
New fasteners are of course fitted… slowly…
Krell KSA 100
Before…
Krell KSA 100
And after. Most never bother with these details. Heck, the last guy snapped the rod and just left it with only three holding the fan and module in place. “She’ll be right!”
Krell KSA 100
Most owners never look at the bottom of their amplifiers either, but I hope the owner of this one at least knows that this is what it looks like!

Bias Discussion

A little technical aside here for those who are interested in the details of standing current, idle current, quiescent current or bias. I’ve added this because there is misinformation out there on the interwebs about this, specifically about this amplifier.

Krell KSA 100
Here, I’m engaged in the final testing and adjustment of this fine amplifier. I’m setting quiescent current in stages and after ramp-up over some time with a variac, to ensure that everything runs smoothly and safely. We don’t want any accidents after all this hard work. Accessing the adjustment points is especially difficult in the KSA 100.
Krell KSA 100
That’s right folks, nearly 600mV across the emitter resistors here, wowsers! This amplifier is acting as a room heater at this point, but she is rock solid, stable, and running perfectly here under test. You’ll see I’m also monitoring a power supply rail voltage.
Krell KSA 100
Why have I set the quiescent current to 575mV? Because that is what is stipulated in the factory service documents, see above. Original service docs are always the most accurate source of information and anyone who even briefly examines my work will understand that accuracy and precision are the foundation of my approach. I offer a best practice approach, better than the factory would offer, in most cases.
Krell KSA 100
So why have I posted in-confidence service details when I am normally against doing so? Well, my article was linked in a Krell forum and, whilst there’s overwhelming praise for my work (thank you), someone at Krell felt that the quiescent current value I’ve used is “wrong”. I couldn’t let this go un fact-checked, also let’s dig.

We all make mistakes and if/when I am wrong, I’ll say so. One can only hope that others apply that same standard… If I’m wrong it means the official service data is wrong because I got the data straight from the original factory service documents. The point was made that these are old documents. Of course they are, it’s a 40-year-old amplifier! You’ll note the 2004 revision date of the document above though. This perfectly correlates with my older 1980s documentation for the Mk 1 and 2. The bias figure I’ve used specifically applies to the post-1983 versions of the KSA 100 and is the only figure that should be used.

Are all the various factory documents quoting this specification wrong? Of course not, and more importantly, they’re the closest thing we have to the original design specifications. Any supposed error would have to be taken up with Dan D’Agostino and until he chips in we have no better source of information for this amplifier. A figure like this is a technical spec, not open to interpretation. Any deviation from it takes us away from the original design intent and amount of power available in class A. Worth noting that when you do the calculations for where the bias should be for this amplifier, we arrive right at 600 mV.

I may not have all the documents and Krell may have revised the quiescent current setting for this amplifier, but halving it is problematic as this will then no longer be a 100W class A amplifier. You can’t have 100W pure class A from this amplifier with less than half of the specified quiescent current. Using a bias figure of 200mV will yield a class A power output of around 40 Watts. I have to assume that the folks at Krell and others in that forum understand this.

If Krell has officially revised the bias spec on these amplifiers there will be documentation explaining the details and implications. Without any supporting docs and based on opinions in a forum rather than hard data, it’s just opinion and that doesn’t help anyone. This is why I seek original sources of technical data wherever possible and the best source for the KSA 100 is the original service data, written by the design engineers. I know what I trust.

So, wrong bias? No and using basic math, the value I’ve used is the only correct value. Having an opinion on this is like having an opinion on the shape of the earth. What nobody wants to see from a company once famous for its engineering is smoke, mirrors and BS. That would be a real shame.

If I discover any updates, I’ll be sure to post them here. I hope someone, somewhere has learned something. I suspect I’ll still be accused of using the ‘wrong’ bias. I wonder how all the KSA 100 owners who’ve recently had Krell overhaul their amplifiers feel about this…
Krell KSA 100
The Krell KSA 100 dissipates a continuous 475 Watts when set per the original factory specifications. That’s more than my Accuphase A-75 and a lot of standing current, but certainly in keeping for a 100 Watt class A amp. This amplifier cannot sit inside a cabinet.
Krell KSA 100
DC offset was initially way out and she’d been running like this for years.
Krell KSA 100
And that’s what we want to see, less than 1mV offset, amazing. This makes a very audible difference by not biasing the speaker cones.
Krell KSA 100
Here are the parts I could find at the end of this job that I’ve replaced. Lots of hardware and some very big capacitors. There were other bits and pieces I couldn’t find or forgot to include here like the 20 or so lids fasteners I replaced. The job was with me for a long time.
img 9685
These venerable Sprague Powerlytic capacitors have given wonderful service. The new caps have a higher C but are still well within the inrush current limits of the diode bridges.
Krell KSA 100
And that is this rather large job completed, my friends.

Results

With this deep service/overhaul/restoration work complete, I ran the amplifier for an extended test and had a good listen to her. Friends and colleagues know that I always say this – I can hear a really good amplifier immediately, even in my workshop, with my drum and live music ravaged ears, through the baby Yamaha NS-10Ms. This Krell KSA 100 is an excellent-sounding amp. It’s obvious within the first 5 seconds of listening.

Krell KSA 100
These new filter capacitors will likely outlast the amplifier. They will transform its sonic performance also.

Thank you to her very patient owner for waiting quite a while for this one. It may or may not be obvious, but work like this is intense and rewarding, but quite draining. There is only so much of it that one can do at a time, especially when solving problems and working on really important pieces like this. For that reason, I can only do a couple of these a year.

Krell KSA 100
She still looks a little tired on the outside, but this Krell KSA 100 really does purr now.

For those wondering how a nicely restored Krell KSA 100 class A amplifier sounds – in a word, s-m-o-o-t-h! There is a warm, fluid sound quality that one associates with class A equipment and which escapes even the better integrated and class AB amplifiers in almost all cases. A big class A power amp, even a bipolar device one like this, tends to sound warm and effortless, with a beguiling smoothness and fluidity that is immediately apparent, even on little monitors as I mentioned.

Krell KSA 100
You may notice I’ve even fitted new lid fasteners. The ones that came with the amplifier were the wrong thread and had been forced in. These new ones fit nicely and slightly lift the general appearance I think.
Krell KSA 100
She’s an angular beast, that’s for sure, but the handles and fan cooling make her much lighter than my Accuphase A-75, which helps me in the shop, because I have to move everything!
Krell KSA 100

I’m confident that this lovely Krell KSA 100 will provide many more years of good service, for her current very happy owner, and maybe even the next one!

Krell KSA 100
Here she is next to another customer’s Krell KSA-100S, the equivalent model from around 10 – 15 years later. You’ll see how things evolved. They are very different internally too. The KSA-100S features sliding bias, so it is not a true class A amplifier.

I hope you’ve enjoyed this article folks and as always, thanks for stopping by. If you appreciate the dozens of hours that go into work and creating articles like this and wish to shout me a drink, you can use the donate button in the footer.

If you’d like me to overhaul or restore your lovely old Krell KSA 100, KSA 150 or any other beautiful legacy Krell product for that matter, don’t hesitate to get in touch. I can’t promise to be able to do more than a couple of big overhauls like this a year, but those I do receive my full attention 🙂


Discover more from LiQUiD AUDiO

Subscribe to get the latest posts sent to your email.

24 thoughts on “Legendary Krell KSA 100 Class A Power Amplifier Restoration”

  1. Hi Robert, thank you and thanks for sharing this, I had not heard, my condolences to the Krell and D’Agostino family and community. I’ve not spoken to anyone there for a couple of years but sad news indeed.

  2. Hi Mike,

    Another fine job. Probably better than new.

    Some sad news to report. I don’t know if you heard, But. Rondi D’Agostino, president of Krell Audio has passed away. Patrick told me she fell down the stairs at home. Krell is of course closed for the time being as she was the main owner if not the only one. This is very sad news as most will agree. I’m not sure if they will reopen. But some say they plan to.

    Robt.

    1. Hi Robert, thank you and thanks for sharing this, I had not heard, my condolences to the Krell and D’Agostino family and community. I’ve not spoken to anyone there for a couple of years but sad news indeed.

      1. Sorry for the double post. I didn’t think the first one went through because of the link.

        Truly sad. I hope Krell does reopen. I heard Rondi had hired a new manager and she was stepping down. But that was before the accident. Krell has been on a shoestring budget for a while. I own a KSA 250S a Duo 175HD, a Studio DAC and a KBL. I’ve owned almost every amp from the KMA 100’s till now.

        You do a great job. Wish you weren’t so far away.

        Best

        Robt.

        1. Thanks Robert and no problem, I’ll delete the copy. Yeah I’ve not contacted Krell since the request for service data I mentioned back in 2023, but it’s a shame what’s happened to the company. They produced some of the most spectacular and outrageous amplifiers and other gear, but like many, the ‘home cinema’ and SACD rabbit holes really hurt them. Those processors the many dead SACD players and some later amplifiers that aren’t a patch on their old ones caused terrible reputational damage unfortunately. I hope they can bounce back.

          1. Krell certainly got out of their comfort zone. Overdriving the lasers in their optical devices didn’t help. Dan has taken a more conservative approach with his new endeavor. Although extremely priced.

                1. We’ve chatted about this previously and whilst Patrick may have told you that, this is not something an engineering-based manufacturer does, for the reasons I’ve previously explained. Lasers are preset and adjusted by their manufacturer (not Krell) to conform to Redbook performance specifications whilst drawing a specific amount of current. There are no “brightness settings” and “turning them up” to max causes them to die within seconds. As before, let’s move on, but thanks for visiting!

      1. All is well.
        I love it, but barely had time to really listen yet.
        Been flat out with local and FIFO work.
        At the moment we mostly use it when we watch TV.

        1. Well that’s good news, keep enjoying it, it’s a beautiful amplifier and rare to find in such excellent condition. You need A-grade partnering equipment though, that will make a world of difference.

  3. Greetings Mike,

    Thank you for the informative article on restoring a KSA-100.

    A short time ago one of the other engineers at work was tasked with cleaning out a store room, and discovered a KSA-100 sitting in there. Luckily, rather than just getting rid of it, she dropped it off in my cube, knowing that: 1) I’m a fan of high quality vintage gear, and 2) I’m particularly interested in power amplifiers (I have some unique designs of my own that I’m in the process of developing.) I have managed to haul the beast home and it’s here in my shop awaiting enough space on my bench to pop it open and see what’s there. It’s in fairly good shape externally except for some missing rubber feet, but at this point its internal condition is still a mystery…

    This one is serial number 524. Do you happen to know when that would date from?

    1. Hi Douglas, my pleasure and I’m glad you found my article interesting. A serial number of 524 is reasonably low but several hundred higher than the one in this article for example, so maybe late 1980s perhaps?

      1. Hi Mike, thanks for responding so quickly!

        After writing my comment above, I couldn’t resist popping the top off and taking a look. This unit is completely original and looks like it has never needed repair! All of the output transistors are labeled “Krell”, and this one has the toroidal power transformers too. There is no sign of any kind of thermal stress and the inside is barely even dusty. I’m willing to bet it has been sitting in that store room for years unused, and just works, but just to be sure I’ll bring it up slowly on a variac so that the power supply caps have a chance to re-form, at least for initial testing. I may do a full recap job on it, but first I’ll find out how close to specs it still is.

        Also of interest are some differences from the one you worked on: this amplifier only has single rectifier bridges on each side, doesn’t have the relay boards, has fuses in series with the main supply rails, and the case design is different.

        This amplfier is what another engineer and former co-worker calls a “cat warmer”, although since this one has the downdraft fans, if a cat sits on top of it, it’ll fill up with cat fur!

        1. No problem and yes, definitely a cat warmer, my cat loves my Accuphase A-75! You don’t want anything blocking air flow on these KSA 100s though, or any other class A amplifier for that matter. Note also the biasing section, current Krell employees are seemingly trying to re-write the technical details without really understanding them it would seem.

  4. This is actually a very early KSA-100 looks like! The earliest models had that shiny metal front (later changed to a matte brushed finish), the fans on top (later placed under the heatsinks, PITA to get to!) and I’m a bit surprised to see EI transformers, all the later ones including MK2’s had toroidal transformers inside.

    1. Hi Martin, this unit was several thousand along the line but yes, still an early one as mentioned in the article. Lots of Krell power amplifiers used EI transformers including various later models, the twin toroidal versions of these are cool though!

Leave a comment, like, share and subscribe!

Scroll to Top

Discover more from LiQUiD AUDiO

Subscribe now to keep reading and get access to the full archive.

Continue reading