Simple HT supply build

[Mike Watterson]

Negative Doubler with its first diode cathode and output cap (1 μF 400V met-poly) connected to OV gives -230V.

OK, I realised how to move one connection and make this a more useful -115V

[Mike Watterson]

I tried that and it's about -100V no load. Drops to about -60V on a 220K Ohms load. So I'll put it back as it was and regulate it with a diode to about 110 to 120V..

I also added a pair of 270K resistors in series as a load on the +700V approx, otherwise nasty! Centre tap was 356V! Eeek! The resistors are rated 200V. I'll move the 0V end to the +350 rail.

[Mike Watterson]

The postman came 2 hours early.

The little panel meters had no instructions. Investigation showed the red and black on 3 pin plug are power and yellow is voltage sense, which can't go to 100 as the DP doesn't move. In theory goes to 99.9V.
The current shunt is a SM resistor which needed desolder braid and sharp knife to remove. Then even with a 100K series resistor it showed about 62.0 if any voltage applied. I added a 10K to zero volts.
11V gives about 52.3 on display, which would be about 1V in.

There are two presets to set V and A calibration.

So an opamp (Norton LM3900, though there is a later upgrade (1981) the LM359. doesn't need a transistor current source added on output. I need to make 50 mA on sense resistor in series with HT be about 1V on op-amp out. I will see if the preset allows setting 1V in (without shunt) to be exactly 50.0 on display.

The Voltmeter part measures 311K when off. At 11V on display it takes 34.7 μA, which is 297K. I'll see if a suitable series resistor assuming 300K allows the voltage adjustment to be exactly 10.0 at 100V, so 700V would be 70..0 as the spec claims it measures to 100V (99.9 really). That would be 2.7 M Ohms? Price was £7.77 in. for TWO on Amazon.

I also got some packs of stupidly cheap LED voltmeters. Three packs, each with 4 displays (red, yellow, green and blue). So TWELVE little LED voltmeters for £8.77 total on Amazon. Also no instructions. The PCB has a solder jumper (not always soldered) to connect Vmeasure to +Vsupply. Opening that and adding a wire works. The displays work off 5V and looks like a LDO regulator. There might be instructions here in German.

I searched the link and there is, and you can select English on the site and for the ebooks,
The little voltmeters.

It's a shop, so they want your email!

This might even be the same V & A module (but I didn't buy it from them)

EDIT:
The Download links appear after you complete checkout, so if you are paranoid you can use a fake email, name and physical address and not save it. Obviously real info for physical purchases.

Of course they are PDFs, not ebooks.

The booklets are pointless if you understand electronic as the connections are obvious from examination. That and specification is all that is in it. The spec is nothing extra than the Amazon web pages.

[Mike Watterson]

The mini 3 digit voltmeters: 12 off for £8.77 on Amazon.

AZDelivery Module Mini Digital Voltmeter Panel 2,5V-30V DC 0.28 Inch with 7 Segment Display Voltmeter 2,5V-30V DC

The Mini Voltmeter allows the reading of the voltages in a simple and fast way.
Due to its characteristics, the Module is perfect to be installed in a house or another structure for a long time.

Since it has Welded Cables the quick test of batteries or accumulators can be developed.
In addition to its versatility, it has an easy connection.
Just connect the black cable to the negative pole of the battery and the red cable to the positive pole of the battery.

The reverse polarity protection makes it safer to use, especially for temporary installations.
In case the reading is very accurate, the voltmeter can be calibrated, through the potentiometer on the back.

The different colours displayed on the screen have different minimum voltages.
This is because different coloured LEDs require different amounts of voltage to illuminate.
So, each colour reflects a different range, the voltmeter with the red screen can measure from 2.5 to 30 V.
The modules with the green screen and the yellow screen work from 2.6-30V and the blue screen from 3-30V.

✔️ Measuring range: 2.5V/2.6V/3.0-30V.
✔️ Operating voltage: Minimum 3V.
✔️ Screen: 0.28" LED.
✔️ Screen colours: Red, Green, Yellow, Blue.
✔️ Measurement accuracy: +- 5%.
✔️ Cable length: approx. 20 cm.
✔️ Dimensions: 31x14x10mm.
✔️ Weight: 4g.

I recommend opening the solder link and a 5V PSU. Spare pad is the voltage measure. I'll check what series resistor they need for x10 = 300V. I'll power one with a floating/isolated 5V supply and connect the 0V connection to -HT via a scale resistor and V sense in to HT supply 0V.

[Mike Watterson]

The combined A & V panel meter is £7.77 inc. for two.

Manufacturer ‎VooGenzek
Part number ‎GZ039T01
Item Weight ‎70 g
Package Dimensions ‎17 x 10.4 x 2.4 cm; 70 g
Manufacturer reference ‎GZ039T01
Size ‎10A

Voltage measuring range: 0-100 V DC; current measuring range: 0-10A (decimal point moves automatically).
Display mode: Dual-colour digital display 0.28" LED digital tube Display color red+blue.
Built-in Shunt: The 10A voltage and current meter itself has a built-in shunt.
Color distinction: Led Display Voltage and current simultaneously in different colors, red for Voltage, Blue for current(the current is measured in the negative rail).
Connection Protection: With reverse connection protection to avoid burning from reverse connection.

I'll recheck if the DP moves. But I don't think it does on amps or volts.

[Mike Watterson]

The small cheap 3 digit voltmeters are about 98K input resistance, so about 882K series for 10:1.

The VA panel meters DO move the DP on current input. It jumps to the left by 1 place below 01.0 on display., so 0.01 to 0.99 is displayed. Later I'ss how well the A and V calibration pots work!

[Mike Watterson]

I tried that and it's about -100V no load. Drops to about -60V on a 220K Ohms load. So I'll put it back as it was and regulate it with a diode to about 110 to 120V..

I also added a pair of 270K resistors in series as a load on the +700V approx, otherwise nasty! Centre tap was 356V! Eeek! The resistors are rated 200V. I'll move the 0V end to the +350 rail.

I put -HT rectifier connection as it was. Added 47 K series and a reversed 1N4148 (randomly chosen) across the 220K resistor that will be replaced by 220K linear pot. That gives a reasonable 96.5V. The 1N4148 are stupidly cheap and seem happy with < 1mA reverse current (that would be < 100 mW). The raw -HT fell to 176V.
The 220K is a load of 440 μA. The 176 - 96 = 80V, thus 47 K Ohms source gives 1.7 mA. That means 1700 - 440 gives about 1.3 mA = 120mW, so the 47 K Ohms is a bit low, or I could use a 100K pot, leaving about 800uA in the diode which is about 80mW, much better.

[Mike Watterson]

Hmm.

You know the cheap USB PSUs? A transformer that would be small in a 1960s transistor driver stage. The ones that you wonder is the CE mark real? It was a cheap 5V 500mA PSU with Europlug style pins that came with some gadget, I forget which.

I powered the 500mA USB PSU off the lowest HT, centre tap of the two 470 μF caps. I connected the V measure to 0V and the common via a 2.7 MΩ resistor to the 100 kΩ pot across -HT 96V to 0V. The USB PSU started smoking at -46V!

I think a floating PSU with genuine 4 kV isolation is needed. Or an LM3900 based level translator that runs off +12V. That would make it easier to do data capture too.

[Mike Watterson]

Also two shots of the stupidly cheap panel meter (Under £4 inc VAT and postage etc).

The currents are fictitious as removed the 10A shunt.  10K across input and 100K in series. Without the 10K it goes to about 62, i.e. full scale.

0.2V
   

and 10.7
   

So volts can be x10 (it pops out of bezel so I can blank Volts decimal point) and current can be mA

Also scary lab bench with over 700V DC at very low power and 350V at maybe up to 20W. Also -145V.

Unless you are experienced, don't do this. Instead assemble my finished design on chassis in a box. There is a slight mitigation that nothing is earthed and it's isolated to 4 kV, so like a bird sitting on HT grid, you are fine only touching ONE place if it's live. It's "live" for a while after power off till the bleed resistors have done their job.

Visible is the small LED voltmeter for -HT and the 100K pot for 0 to -95V. I think I'll add a 1/4 LM3900 or similar to change the voltage to a 1/10th positive version.

Weird filenames as I used Viber chat on WiFi on phone to send photos to workstation Viber "My Notes".

[Mike Watterson]

Next I'll find a box and mount the scary raw HT parts in it.
Then
1. The op-amp HT current supply side sense to give 1V out for 10mA series current through 10 or 100 ohms. Also to translate 0 to -100V to 0 to +10V
2. Put all that and pots, switches etc on / in box with leakage flash neon and second mini-LCD meter to show breakdown voltage. Add 5V supply for meters and 12V supply for FET bias. I might derive a regulated -12V or -15V as an addition or alternative to the -95 V nominal.
3. Maybe a variable + Bias supply using an LM317 (testing HV FETs, or valves with very low HT).
4. Add lamps and FETs on heatsinks
5. Final test and changes.

Note if you feed the FET gate bias pot from a scope X timebase ramp you can have a simple curve tracer. Put the 1V drive for current meters to scope Y1 and HT of Anode or g2/screen to Y2. Or whatever. Or have a ramp generator at 50Hz and feed scope X with HT.
Maybe want to add varying the -HT that supplies grid. That's slightly more complicated. A neon and capacitor would be a simple ramp generator and could be offset.

So maybe need more switches and conectors. Use 4mm wander sockets or BNC?

Not decided yet if I'll add another +HT breakdown/Leakage multipler stage. Any opinions..