Page 1 of 1

FAQ - Build a high voltage "bias" meter for free!

Posted: Wed May 15, 2019 8:26 pm
by Richard Hull
Bias voltages, PMT and GM supply voltages often need to be measured with accuracy. A bias supply is an inherently weak supply. Portable instruments that are battery powered typically have extremely weak bias supplies. (capable of just a few microamps due to their Hi-Z load characteristics. It is rare to require any nuclear detector bias much above 2kv.

***************************** SAFETY NOTE *********************************************************
The meter described and constructed, as below, cannot be expanded to higher voltages! This is due to the mounting of the divider chain sealed in shrink tubing and epoxied to the meter case. This meter is a bias meter and not to be used on high current 2.5kv voltage tests.
DO NOT hold the meter in your hands while making any high voltage measurement.

In some repair and maintenance manuals, it is recommend measuring the high voltage with an electrostatic voltmeter!!! Do you have one? The average DVM has an input impedance of 1 megohm to a maximum of 10 megohms. This will load a bias supply to the point of having you set your voltage and the instant you remove the meter, the voltage will skyrocket. Your meter loaded it down, you adjusted to the recommended level and now it is 20% higher in reality!

I will give a simple way to make a free 100 megohm digital voltmeter. It is abysmally simple to any better than average electronics guy on how to do this, but here is a method that is cheap and easy for anyone.

1 Get a free Harbor Freight DVM.....These are the list priced $5.99 meters. They have several types. Get the low end model. There is no shame in shelling out $5.99, but this is about a free meter, not a $5.99 plus meter. They do often have coupons for these meters for FREE in their store flyers, local Sunday newspapers, American Rifleman and other manly men's magazines! Once you give them your name and address at the register, you will get "free stuff flyers" in the mail, too.

Note* these little Chi-Com meters can be stunningly accurate! Using a Weston "standard cell", I have found they are often accurate to within +/- .001 volt. WOW! (Weston is 1.019 or there abouts based on temp) One megohm is a rather heavy load on the cell. Every old Fluke 50s-70's, hyper accurate differential voltmeter had a standard cell in it. A well made "standard cell" is usually good for a couple of hundred years!! 1.019 volts forever!

2. It is hoped you have a small supply of high ohm resistors around in the 10 meg and above range.

I supply an image below that is more or less self explanatory for those capable of monkey see, monkey do.

The epitome of the process..........................

The Harbor Freight meter has an input impedance on all ranges of 1 megohm. We need a precision matched 99 megohm resistor hooked in series with the red positive meter lead. Once this is done, we put the meter on the 20 volt range and now have a 0-2kv 100megohm impedance meter that will not load even the weakest of bias supplies. The best thing we can do is to get our hands on a stiff, (low-Z), DC supply of about 300-900 volts. (you are on your own here. Like maybe make one up with a transformer variac and some diodes and a filter capacitor with suitable bleeder!)

Using the Harbor Freight meter, as purchased, on the 1000 volt range, adjust the supply or the variac controlling it to a voltage at or above 300 volts...Precisely!!!! The voltage is unimportant just make it an even voltage at hundreds. (300, 500, 700, whatever) Turn the supply off. For the following text, we assume you have set to 500 volts. Turn off the supply!!! (note edit may 2019)....The current spate of harbor freight meters no longer sport a 1000 volt dc top end range...too bad.
get you hands on any dvm with a 1000vdc top range.

Start assembling your 99 meg resistor. I used (5) 18 meg, 1/4 watt 5% film resistors in series to start with. Hook this partially done, series resistor in series with the positive probe of the Harbor Freight DVM and the other side of the resistor to the hot of the HV supply. Needless to say, you hook the black negative probe to the negative of the supply. Place the meter on the 20 volt range. Turn on the supply. You should read a good bit high on the voltage. If you set the supply for 500 volts, you might read 5.45 = 545 volts. Now, to avoid killing yourself, TURN OFF THE SUPPLY!! If high, You need more resistance. Try a 3 or 8 meg resistor added in series. Retry and add resistance as needed until you are between 4.98 and 5.02. You are done! If you read low like, say, 488 volts. you will need to take some resistance out of the series string. In the end, you will have an ultra high impedance (~100 megohm), 2kv meter capable of better than 1% accuracy.

You can fancy this thing up in any number of ways you like. As I have over 54 of these free meters collected over the last 6 years, I dedicated my meter to a bias meter. (see the photo) I think I wound up with 7 resistors in my series string. Since I opt'd for boring a hole in the meter and connecting one end of the resistor string to the positive input on the meter's circuit board and planned on epoxying the string to the meter's right side, I shrink wrapped the string to avoid shock if holding the meter during testing. (see photo) I fashioned a little clip loop at the top end of the string for an alligator clip lead when in need of bias adjustment. By doing this, the meter is still good for use as a normal DVM with the special lug on the string for the 2kv 100 megohm bias setting. Just remember to put the meter on the 20 volt range when doing bias adjustments or setting GM or PMT supply voltages.

I do a lot of development and custom work on battery operated weak, (hi-Z), supplies in this range and need to know my voltages under load without further loading the supply while testing or making in circuit adjustments.

Good luck if you plan on making this bias meter.

Note*.....Try and use the highest voltage under 1000 volts to build your string. I used 800 volts. If you use 200 volts and settle for a reading of 203 on your string, at 800 volts you will be at 812 volts and at 1600 volts you will be at 1624. In short, the error multiplies. I settled at 801-2 bobbling. As a rule, HV bias levels need not be precise to the volt, but 1% or 2% accuracy is desirable

Richard Hull


Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Sun Jun 30, 2019 11:35 am
by Cai Arcos

The low-Z high voltage source proposed by Richard for calibration can be dangerous, specially due to the capacitors needed to reduce ripple. I have devised a method for calibration that doesn't need such a supply.

First, connect a standard (10 or 100 mega ohm input impedance) multimeter to the output of the high-Z supply. This will load the supply, and decrease the output voltage. However, as long as it is connected, this voltage reading will be accurate. Now, do the resistor-adding-substraction procedure described above until the reading on the chi-com multimeter matches that of the standard multimeter. If you know remove the 10Meg multimeter, the voltage will skyrocket, which will be visible on your new high-Z meter.

This way, the same nuclear detection supply can be used, which is a lot safer to handle. Furthermore, you might be able to achieve a higher voltage for calibration than that possible to obtain using a simple variac.

Cai Arcos

Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Sun Jun 30, 2019 9:08 pm
by Richard Hull
Great idea Cai! The average NIM HV bias supply is capable of far more current than I made my meter for, (as much as 5 ma). My meter was built for GM work which needs only 50 ua, if that. I make up tiny supplies for portable gear. Space and current drain on battery operation must remain very low. My supplies typically drain the battery which is regulated to 5Volts to no more than 15ma or an average of 75 milliwatts at 700 volts that allows for a continuous current with 90% efficiency of .068W = I^2 X 10e7 = ~80ua max current into the 10meg GM resistor.

The 100meg meter impedance here in my meter only demands about 8ua. At this loading, the meter can read true.

I would not rate this type of cheap meter to read much beyond a stable DC reading of 5kv. As the scaling on all such 200mv basic displays limits the display to factors of 2, I chose 2kv for my upper limit and this will handle most PMT, GM, 3He neutron detector supplies and even some small BF3 tube supply needs in portable applications.

The value of portable, battery powered instrumentation can go far beyond use in the field and move into the lab. This allows for a lot of noise immunity from common grounding failures on the AC line provided superlative RF shielding is part of the portable design.

Richard Hull

Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Mon Jul 01, 2019 6:33 pm
by Cai Arcos

I have made one final refinement to the bias meter.

The resistor adding and substracting procedure can be a bit tedious, and the accuracy obtained in the higher ranges is always a bit lackluster. Furthermore, if you are like me (your local electronics supplier is not very well stocked and don't like to wait for shipping) obtaining a wide range of high value resistors can be a bit difficult.

What I've done is taken simultaneous measurements in the bias meter and in a trusted one (as described in my last message). I then constructed an Excel table containing the "Real" value (the value measured by the trusted multimeter), the "Measured" value (the value obtained with the bias meter) and the "Error" (substracting the real measurement from the measured one).

As Richard has stated previously, the error grows exponentially. So I take the measured values (¡the ones measured with the bias meter, not the trusted meter!) and the error and using Geogebra (because apparently Excel can not make exponential best fine lines), created a best-fit exponential line equation (with y for the error and x for the measured voltage). This equation now provides me with the error which I can readily calculate from the measured value and add it to itselft to obtain the real value. In my case, and to make it easier, I have wrote this equation on a piece of paper and sticky tape it to the back of the cheapo.

This way, a little bit of mathematics can assure a much higher accuracy (specially in the higher ranges) and make the calibration easier.

Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Tue Jul 02, 2019 10:49 am
by Richard Hull
NEWS....Harbor Freight, now, rarely offers these meters free via coupons any more. ( I used to get 5 or more free meters with 1000volt ranges per month!). Ah the good old days....

Their current top voltage range is only 250 volts (household voltages). Still fine for a bias meter on 20 volt range.

The purchase price is now $6.95 instead of $5.95 (Tariffs on Chicom goods??)
"Less is more" (1984 Orwellian-speak)

Richard Hull

Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Wed Jul 03, 2019 9:59 pm
by Bob Reite
I've been looking for those free meters, the only thing I see for free now from Harbor Freight are the LED flashlights.

Re: FAQ - Build a high voltage "bias" meter for free!

Posted: Thu Jul 04, 2019 9:02 pm
by Richard Hull
Yes, as I noted, they are rare now. I still find them on occasion in their full page ads in my "American Rifleman" magazine. I haven't see them offered in "val-pak" mailings or in the newspaper Harbor Freight flyers either. Perhaps, due to tariffs, they are just beyond the freebie range now. I currently own over 141 of the meters and never paid for one. (collected them constantly since 2012) About 25 of them are the newer 250 volt max meters. Everything Chicom just got a bit more expensive.

Currently, I have seen these for about $6.00 - $7.00 at Harbor Freight.

Richard Hull