This is a PM6303 Philips RLC Bridge.
My original intent was write about a Tek 2215 that I'm restoring. I was about to buy the time/div switch necessary to finish it when I found this poor bridge on a auctions site by the same price. I just couldn't resist.
It was in really really in bad shape.
As we can see in the picture, it suffered some kind of mechanical damage. I suspect it was just thrown in a scrap pile with other heavy objects thrown over it. When I tried to gain access to the interior of instrument, the case was so damaged that I had to remove the rivets with a drill.
When opened, the things got worst.
Não foi fornecido texto alternativo para esta imagem
It had signals of contact with water (rain, I suspect). The main board had trails of dust following what appears to be the water flow path. Of course, I had to wash the board before debug it.
With every thing washed and dried, lets turn on and verify the damages…
...but first, a few lines about this instrument.
As I could learn about, this bridge was built somewhere near the acquisition of Test & Measurement Philips business by Fluke. And it was a really honest piece of engineering.
It uses the I-V converter method to determine an DUT impedance. Its microcontrolled, and uses a MCS-48 family processor. The MCU by itself is not impressive and don’t have even a AD converter. Almost all heavy work is done with a circuit built with relatively common components. There are a 16MHz square wave crystal oscillator, a series of dividers to obtain the 1KHz test signal and an active filter to turn it in a sinusoidal wave (the excitation signal).
Não foi fornecido texto alternativo para esta imagem
The excitation signal is applied to the DUT and CMOS analog switches select from the voltage divider formed by the DUT and the bridge circuit, including a shunt resistor from the I-V circuit. This signal is applied to controlled gain amplifiers, made with LF356 op amps and more analog switches (4052 and 4053). The resulting signal is applied to a phase controlled rectifier, that selectively rectifies the signal in a sequence of wave angles (0, 90 & 180). The rectified signal, feeds an integrator, that generates a time mark according to the voltage value present in the specific wave angle acquired. The time mark stops a counter formed by TTL ICs and trigger the action of MCU. Its basically a voltage to frequency converter that can select (and limit) the angle of acquired wave.
From the circuits theory, a inductive, capacitive or resistive DUT will have distinct behavior in therms of current and voltage waveform angles when excited by alternated signal in that circuit. So, analyzing voltage values at specific quadrants of the sinusoid, we can tell if we are measuring a R, L or C. And here, the MCU does its magic, calculate that values and show us its meanings and magnitudes.
Math done, the processor sends the result to an nice LCD display thru an interesting Mullard/Philips chip, that is basically an shift register with LCD compatible outputs (provide the kind of alternated signal needed to make the display work). What kind of results to send? A LED signaled menu tells to the MCU what the user want. A default condition lets the processor decide what appears to be the correct. Of course, the user must interact in some cases, hence the other options in menu.
Beautiful, simple and well done. It doesn't demand a lot from the processor, doesn't try to be too bold on analog.
But lets come back to my unit…
Instrument ready for test, I turned on… and (almost) nothing.
The display was dead. The LEDs menu was active, except by the default position. No signal of activity. Lets start…
The first thing I noticed was the silence in the MCU bus. It was hot, but doesn’t even was accessing the EPROM. It is a 8039. Searching my parts stock, I didn’t find an exact replacement, but found two EPROM windowed and one OTP version of 8749. The OTP version was there just because I forgot to throw it away. I had disassembled some thing and don’t checked the parts saved before store it. But like in the MCS-51 family, the difference between a MCU with an internal ROM and a MCU without it is just a enable in the right pin.
I tried a replacement with the OTP version and it worked! The display was back.
Não foi fornecido texto alternativo para esta imagem
But I had no reasons to celebrate yet. The read values was a random variation. There was more to be done. And checking for the excitation signal, I found a open resistor before a clamp made with a pair of zeners. First suspect: somebody tried to measure a charged capacitor.
I followed the signal path and found a lot of diode clamper circuits with shorted diodes. After diodes, the op amps in contact with the DUT was dead, too. Same thing with the analog switches… and same thing with two analog power signals of +5V and -5V, that came from simple zener regulators.
I replaced a lot of these chips, installing DIP sockets in their places to ease my work. Same thing with the open and shorted components. Checking the inert LED, it was mechanically broken. One more replacement.
Turn on and test a known DUT… and its back to life again! Or almost. It doesn’t work with all values of DUT.
After some testing I discovered that the analog power signal +5V was producing an abnormal ripple wen a low Z DUT was tested. Checking the zener specs, I discovered that my replacement polarizes with more current than the original. I make new calculations and added a parallel resistor to the original one to give more current for the circuit.
And now it works!
My Philips PM6303 is back to your old glory!
Or almost. While the electronics section is now fine, I don’t reserved a fair budget to the mechanical restoration (and I’m afraid I lack of the necessary talent to do it). Not sure of I can call its present condition ‘fair’.
P.S.: If you know the Fluke PM6303, compare it schematics with the Philips PM6303. Its the same instrument with a new processor and display.
Home
Help
Search
Login
Register
