Bruel & Kjaer 2606 measuring amp repair experience

[1audio]

I picked up this older measuring amp off eBay recently. I have been assembling a microphone calibration setup and getting the pieces to properly do insert calibration has been difficult. I was even exploring modifying an existing measuring amp or power supply but the mechanics of these boxes makes that quite difficult. In any case this came up on eBay and supports external insert voltage so its well suited to my specific need.

B&K made several variations on the type of measuring amp The 2606 (this one) and the 2607 were from the late 1960's. They they introduced the 2610 and 2636. The newer ones do not support external insert voltage so not suitable to my needs. The 2607 has a log or linear meter response. however you need the collection of meter scales to get utility from he feature. B&K is to my knowledge the only company to embrace user swappable meter scales to enhance the products.

One peculiarity of these measuring amps is an input and an output attenuator that actually change the gain of the stages. The manual is not too clear on why with some suggestion about watching for overload. There is a logic system syncing the two switches and displaying the total scaling on lights across the bottom of the meter. I have not figured out why or how best to use this additional complexity.

In any case this was dead on arrival so time to figure out what's wrong. I traced the signal to a point that turned out to be a board marked "bad". No extension board and limited documentation plus difficult access left me shotgunning and adding wires for test points.  After changing some transistors with no effect i discovered a vintage polystyrene cap was shorted (new experience for me). But I discovered the transistors I replaced were interfering with the range switch, displacing the board and causing an open circuit at some levels. Discovering the the PCB's were part of the range switch AND removable was a first, having never seen such a clever solution before. Next problem was that the response was off by a lot. And looking through it there are a lot of small electrolytic caps. Some seem fine despite 50 years of service. And some are distressed. It turns out that a particular 2 uF electrolytic used in a lot of places had all basically become open.  Small axial electrolytic caps are not used much at all any more but in the late 1960's
were the choice it seems. So I needed to get a supply in to proceed.

Next surprise was that this generation B&K uses the older style preamp connector which is approx 2” long. The new ones are only 1” and need an adapter to work in the older boxes.  (see picture). This along with the idiosyncratic signal connector B&K used makes these older instrument a pain to integrate with. Fortunately the B&K preamp with insert voltage also has the early style connector and works in this instrument. It also has an odd trimpot with a lock screw. You loosen the screw and manually slide the tap and then tighten the screw.

It all seems to be working now so time to integrate with my GR 1559 reciprocity testers.

[1audio]

Update- the instrument is all working HOWEVER something in the RMS-meter driver is not working right. The meter tracking is screwed up. Its quite non-linear. with the middle stretched out at both near zero and full scale, Its really frustrating sine the docs are devoid of info on adjustments and calibration of that circuitry.

I have attached the RMS circuit. It seems to use sequential transistor addition to create RMS from the sine/other wave. I don't really follow it. All the transistors seem to be functioning but at this point I may shotgun replace the lot. I'm open to any useful suggestions.

[1audio]

After much anguish and shotgunning I found the tracking problem in the RMS converter circuit. I replaced all the transistors on the RMS circuit with no change. And then I realized the problem was a fixed offset of around .7V or so. The 44 dB amp in front of the RMS circuit has a pair of diodes set up to create a crossover distortion bump intended to mimic the errors from the bridge rectifier. It seems the original diodes were not creating enough of the .6? volt step needed to make this work. As a guess I added 1n4148's in series with the existing diodes to get a larger step. It worked.

Also the amp is very sensitive to leakage in some of the caps. The caps in this a quite old so I replaced some only to have the amp circuits slowly fad out after power up. I finally determined that some caps were leaking enough to cause DC shifts and the output going to the positive rail. Better caps are on order. Unfortunately instruments of this vintage used axial electrolytics which are mostly obsolete. I'm not thrilled to put 30 year old caps into something as complex as this so finding suitable caps is a hunt.