Differential Current Balance
The Diff Mod. There’s quite a bit about this on our site, but I still get a lot of questions, so here’s a page about it.
Audio amplifiers have steadily evolved since the late 40’s/early 50’s when this whole hifi thing got started. Of course, in the beginning, there were tubes. After a lot of evolution, they sounded really good, but had a lot of shortcomings. Like… poor bass definition, heat, relatively short life, low power output, and the requirement for heavy and very expensive transformers.
When the transistor came into more common usage, transistorized hifi and then stereo amps began to appear on the scene, in the early to mid 60’s. In the beginning, they sounded almost universally awful.
This was because the early designers had tried to use the same architecture (circuit design) as the tube amps, without the output transformer, of course. The characteristic of tube amps was that they were really low distortion for low power levels with the distortion getting worse with higher power. The transistor amp was opposite, with high distortion at low levels, getting cleaner at higher power levels.
The transistor amps sounded really bad, because all the details, imaging, character, of music was in the low level stuff. You could kinda think of a tube amp as doing really well on mid and high frequencies, and transistors doing well for bass.
One of the reasons for this is that the transformer itself along with the circuitry tended to distort with even numbered harmonics. Transistors tended to distort with odd numbered harmonics. The even harmonic distortion is very similar to what a musical instrument does, so the tube amps sounded better.
In 1967 that all changed when Crown came out with its DC300A amplifier. This amplifier used a different architecture. They used a circuit not generally used in tube amplifiers called the Laboratory Amplifier. The lab amp in tube form had been around for years and was mainly used in laboratory test equipment like Hewlett Packard Oscilloscopes and VTVMs (vacuum tube volt meter). It was very accurate and precise. Crown called their new amplifier the Laboratory Amplifier.
It was such an improvement over the earlier transistor amps, and offered so much power, (300 watts per channel) that audiophiles bought them in droves. This amp also revolutionized the PA industry, because they made enough power (finally) so that a rock n roll band could be heard over 50,000 screaming kids. It is an absolute truth that the DC300A created the Rock n Roll PA industry.
Crown used an integrated circuit called an op amp (the LM709), as part of the circuit architecture. Other stereo amp manufactures copied this design, but left out the op amp. In doing so, they left out an important part of the lab amp circuit. This original op amp was so poor by today’s standards that the other manufacturer’s amplifiers sounded about as good as the Crown, with a lot less complexity.
So, from about 1971 on, as the decade of the Golden Age of Stereo progressed, Japan Inc. continued to evolve the transistor amp, and practically all the manufacturers used this half lab amp design. I’m not sure if the engineers understood what they were doing, ie. leaving out a part of the design, or were just in a big round robin of copying each other. The designs and power increases came out so fast and furious that you had to wonder what was going on.
Toward the end of the decade, some of the manufacturers started to incorporate the entire lab amp circuit into their amplifiers, using various circuitry to do so. You can pretty much tell which manufacturers, and which units had this improvement by looking at the specs. Distortion figures had been running about 0.1-0.5% through the 70’s. The new (complete lab amp) designs had distortion figures an order of magnitude better, or more. 0.01 % or so.
There was an improvement in the sound quality, but detractors talked about too much negative feedback, and transient IM distortion, and these things were very real. The promise of the lab amp was still not being realized.
Now here is the thing about consumer electronics. The holy grail of manufacturing is that you make stuff on an assembly line that requires as little technical adjustment as possible. A unit comes off the assembly line, and you plug it in and it works. Amps already needed bias and offset adjusted by a tech, but by the end of the decade the engineers just about had worked it out so that adjustment was almost automatic.
The original lab amp, as used in tube test equipment had to have each input adjusted for something called current balance. This was a tedious adjustment, requiring several pieces of test gear to get it right.The manufacturers that were making the newer, low distortion designs achieved the goal of “no adjustment required” by using another circuit to do the current balance. In my opinion, it was this added circuit that was creating the “transient IM” distortion and other negative qualities that these amps had.
Now I didn’t think any of this up. I wasn’t even aware of all this at the time. I had been designing all sorts of audio gear, from electronic crossovers to power distribution systems, without ever paying much attention to amplifiers. I got interested in the late 00’s and ended up studying several books on amplifier design by Doug Self. He explains all this stuff (not the history so much) and all the various ways to achieve super low distortion in an amplifier.
Doug’s best way of achieving a complete lab amp design used a circuit called the current mirror. This circuit achieved the current balance without any adjustments, but in my opinion added a sound of its own to the amplifier, that cold but accurate, fatiguing sound that audiophiles talk about and the later super low distortion amps seem to have. What we wanted was warm and musical, non fatiguing, accuracy.
I loved the old vintage gear from the 70’s. I was in my 20’s during that time. After reading Doug’s books, I had this idea… why not add back the missing part of the lab amp to this vintage gear, and do it just like it was originally done for test gear, with a tedious adjustment required… no additional trick circuitry with its own sound problems. I wasn’t going to do anything on an assembly line, it was just for my own satisfaction and I didn’t care if it took me a couple hours to get it adjusted.
That changed though, when I got a couple requests to restore and upgrade 9001’s for some folks. I had helped a guy in Singapore, on line, rebuild a 9001, and it grew to a very long thread on Audiokarma, with a lot of watchers.
The diff mod was born. I struggled for a year or more with circuit design, but then discovered how to make the current part of the circuit addition with an FET. I originally did the balance with racks full of 1% resistors that I would mix and match down to about a 0.1% tolerance in the circuit. I then figured out how to drop a 20 turn trim pot right onto the existing amp driver board, which made it a lot easier to adjust. It still required 3 meters and two different adjustments for each amp channel to do it, but no more measuring, mixing, and matching of 1% resistors.
So, we offer this modification and we call it the “differential current balance modification”. Most people who have us restore their old Sansuis get this mod when we do a restoration.
What does it sound like? Well… it is immediately obvious to most people. My own experience was almost life changing in its intensity, having been in the music industry for most of my adult life.
I first listened to several old Beatles recordings that I had listened to a thousand times. On the White Album, I heard stuff, little riffs and lines, that I had never heard before. I could listen for a long time without getting fatigued. Wow, I really liked it.
You can read some other descriptions on our feedback page.
We’ve only had one person who did not like the results, and I changed that amp back to original, as we say in the guarantee.
An interesting aside to all this is how amp evolution occurred at Sansui. I can only guess from their products, but Sansui put a lot of effort into developing the earlier amp architecture. They made both designs all the way into about 1974 or so. Several receivers and early amps had the pre-Crown architecture, and they made it sound very good.
What’s interesting is that these earlier designs, which are in the AU505, AU555, the early quads, and several others, are very musical sounding. That design requires a large electrolytic capacitor on the amplifier output. Electrolytic caps have improved 100 fold since the 70’s, and if you take a big modern cap (10-20 times larger than originally used) and replace the output cap with it, the sound of these amps can be stunning. The AU505 is my favorite of these, and when upgraded, is true audiophile quality, very musical and non-fatiguing.
So that’s the story of the Differential Current Balance Modification. If you stayed with me all the way through to the end here, I assume I achieved my goal of explaining it in layman’s terms. I tried my best to make it understandable, without getting really technical. If you are a technical type, and are really interested in this stuff, I would suggest reading some of Doug Self’s books. Like I said, I didn’t figure any of it out, it’s all there in his books.