I decided to go ahead and publish what I had ready in a blog post until it is ready for its own page. The AA-51 is what is holding me up on publishing my findings in the BlackGate™ Article! Some of the requirements for calibration aren’t easy and I’m having issues with getting the unit to work in SigCom Chassis mode (vs. float). It’s been a myriad of problems and I wanted to keep my readers informed on what is happening. Don’t think I’m just sitting around forgetting the dates I give myself on when I expect to finish, haha.
Over the past few weeks I have been getting much more familiar with the Potomac AA-51 (and it’s owner’s manual) as I attempt to calibrate the unit. Calibration is essential for accurate THD, Phase Ratio, IMD and other measurements, ones I need for the article. There are requirements for calibration that aren’t exactly easy to produce without a heads up. I’ll cover some of the calibration steps in this article, but you can find the complete steps in the AA-51 manual’s maintenance section (here).
One of the hardest requirements for the average user to meet for calibration is the 1000Hz 80Vrms signal (Signal Leveler section). Most modern DDS signal generators can’t give you much more than 15-30Vrms, so you’ll need an amplifier to achieve values anywhere near 80Vrms. Audio amplifiers can work, but ensure you have a clean signal if possible. The easiest way to check for clipping is with an oscilloscope.
The Potomac AA-51:
(right click and hit view image or view in new tab to see this or any image on HL in full-size)
One of the first calibration requirements to poke its head was the need for a signal generator with dual isolated outputs. The last part is what put a wrench into my plans, as neither of the 3 generators I had in my possession have isolated outputs (they all share a common chassis ground on the BNC shield). A common thought is to simply float the ground, which is a nice way of saying cut them free from the chassis. Doing so in some countries is literally illegal, but this could work depending on your generator. (Safety Note: removing the ground in the wrong situation can lead to a fire!) I’m not going to recommend it, but you can look it up yourself for more info. After doing some reading on how to best remedy this situation, the most likely choice is a simple 1:1 isolation transformer, but it can also be done for less money using capacitors.
Note that if you use multiple waveform generators, you’ll need a way of synchronizing them, usually accomplished through an external trigger, but some don’t have this. This becomes more apparent when you are asked to create a phase change of 45 degrees, this can be accomplished either through the use of the correct capacitor/s (dependent on your wave gen’s output impedance) or you can use a direct delay in the generator, if yours has that option. I was hoping to use my Digilent Analog Discovery2 to do this, since it can give easy and precise phase delays using their easy to understand GUI. Unfortunately they also lack isolation between the outputs!
I attempted to use the LJM MX50SE DIY amplifier to achieve this 80Vrms signal. I was using a 500VA toroidal transformer, along with a large capacitance bank for smoothing/filtering the rectified AC signal, getting closer and closer to true DC (remember this is the power input for the amplifier, not the input signal, which is AC). I had about a 60Vrms signal displayed on my Tenma 72-1020 TrueRMS 40k count multi-meter and then “pop” + FIRE on one of the small resistors!
I am usually honest about my mistakes, as my readers surely know by now, I even blew it up a 2nd time! In my opinion, it’s hard to give advice on something if you have never done it yourself. This makes failure for the author as valuable to the reader as success, perhaps more valuable. When I went looking for help I ended up on an excellent LJM MX50SE thread (found here). There I learned about the problem. I wasn’t positive why one of the 47 ohm resistors went up in flames. I learned that my load was the issue, I was trying to use roughly an 8 ohm dummy load (3x 30W Dale HL-35-08Z 25 ohm resistors in parallel). To create an accurate speaker dummy load you either need to use capacitors or inductors, depending on your requirements (or a speaker, lightbulb, etc).
In some aspects inductors are odd little components, depending on the frequency of the signal (Note: DC = 0 Hz) they can act like a capacitor or like an inductor (in alternating current (AC) circuits)! This is due to how AC signals behave with regard to impedance. For a safe dummy load we need real and imaginary impedances for a balanced load. Most amplifiers would kick into a safety mode, but for an amplifier as bare-bones as the LJM MX50SE, we don’t have that luxury. The weakest link in the chain was the 47 ohm resistor, because this resistor had the most current flowing through it (least resistance resistor in circuit) and thus was the first to go boom
As you can probably tell from the PCB construction and some component choices, my AA-51 was made around the late to mid 1980s. The Sprague Atom capacitors in the power supply section (A1) are dated the 47th week of 1985.
The unit uses Switchcraft 05CL3M connectors which are slightly modified XLR connectors, they have a locking ring/collar to ensure quality connections. Even though the unit is capable of handling 3 conductors per channel, most of the test recommend using unbalanced/RCA signals instead of XLR. In particular the manual says that the THD measurements should be made with unbalanced lines for accuracy. It also says the SigCom selector must be in chassis mode, instead of float. SigCom is another word for the signal ground.
SigCom is my other primary problem right now. Anytime I switch the unit to SigCom Chassis my signal level becomes almost non-existent, including in voltmeter mode. I’ve tried using the dedicated ground terminal on the back of the unit with no success. The other mode for SigCom is float, this is what I have to use (as the other mode doesn’t work, yet). If anyone has knowledge of the AA-51 and the SigCom setting, please comment below or contact me (here).
From my experience, the Sprague Atom™ capacitors are some of the most reliable electrolytics ever made. Most capacitors are doing good meeting the 20 year rule, while 90% or more of the Atoms I’ve checked over the years (hundreds) still read better than most modern equivalents! I’m talking about their ESR, Q-Factor, Insulation/Dielectric Resistance or DC-Resistance, and Capacitance accuracy.
You can date most Sprague capacitors very easily, the Atom™ date code is listed directly on the capacitor, last two of the year and week of the year (0-52). These Atom™ capacitors above are from the 47th week of 1985! Still, when doing the power supply calibration, all of the numbers were within 0.1V of their target values.
(Below you will find each of the removable PCBs found in the AA-51’s socket bay. Each board has a very specific function, details are listed above their pictures.)
PCB A2 – Signal Leveler Board:
PCB A3 – Phase and Ratio Detector:
PCB A4 – Quadrature Detector:
PCB A5 – IM / Wow & Flutter Demodulator:
Definitions/Glossary:
- DC : direct current (0Hz frequency)
- AC : Alternating Current
- 20 Year Rule: A general guideline as to the lifespan of electrolyric capacitors.
- ESR : Equivalent Series Resistance
- DC-R : DC Resistance (Dielectric Impedance to DC Voltage)
- Q-Factor : Quality Factor (Inverse of the Dielectric/Insulation Dissipation Factor)
- Dielectric : The semi-conductor layer in between the capacitor conducting plates, formerly called the insulation, as seen in the term Insulation Resistance. This can be anything from air to Mylar Film™ or Teflon™.
- DDS : Direct Digital (Frequency) Synthesis
- Vrms : Volt Root Mean Square