UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Components and Circuits (https://www.vintage-radio.net/forum/forumdisplay.php?f=40)
-   -   High side high voltage MOSFET switching with 3V logic drive? (https://www.vintage-radio.net/forum/showthread.php?t=153370)

Martin G7MRV 22nd Jan 2019 9:11 pm

High side high voltage MOSFET switching with 3V logic drive?
 
Hi All,

Im working on a project to give modern microprocessor control to vintage 1980s military radio equipment. Due to the majority of the control system being interfaced with 3V logic, I find myself with a puzzle,

I need to use 3V logic outputs, from MCP23017 I2C port expanders, to switch between 18 different high voltage lines generated from a potential divider stack. This is 18 levels between 0 and +80V, currently selected by a 18way wafer switch.

Im thinking, I need p-channel MOSFETs to do the high side switch, but because these need biasing from the HT line to the gate, clearly then need another transistor, either an n-channel MOSFET or NPN, to control it from the 3V logic?

Would I be correct in thinking then that the drive transistor would also need to be rated for the high voltage on the line its driving the switch for?

This is a moderate HT but very little current, and also very very limited space in the radio, so I need small devices as well.

Can anyone suggest suitable devices, that are also low cost? If im looking at 36 MOSFETs to switch 18 lines, I need compact and cheap!

I could of course use different voltage ratings depending on the line being switches, which should allow me to use cheaper parts where the high rating isnt required.

The circuit im working on can be seen in this post on my blog http://g7mrv.blogspot.com/2019/01/ht-chain-issues.html

I'd be pleased to hear any suggestions as to circuits, devices etc, especially if anyone knows an IC that could do the task for me!

Regards
Martin G7MRV

Argus25 22nd Jan 2019 10:58 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

Originally Posted by Martin G7MRV (Post 1113364)
I need to use 3V logic outputs, from MCP23017 I2C port expanders, to switch between 18 different high voltage lines generated from a potential divider stack. This is 18 levels between 0 and +80V, currently selected by a 18way wafer switch.

Would I be correct in thinking then that the drive transistor would also need to be rated for the high voltage on the line its driving the switch for?

Yes, normally the driver transistor (or device) has to have a collector -base voltage rating at least equal to the supply voltage preferably more. If you are using a high side mosfet then most likely you would also use a zener diode to limit the max gate-source voltage to around 12 to 15V. For the driver transistors the MPSA series come to mind but I would have to look up the specs.

You didn't say how stiff the voltage levels needed to be (the internal resistance of the created voltage supply). Also that you need 18 levels, that presumably need to be isolated from each other so they can pass through the selector switch. And you are starting with digital levels. I cannot see the need for 18 mosfets. Will more than one line be used at once ? It doesn't seem so if passed through a selector switch.

If you are starting with digital signals, one way to do this would be to encode that into 5 bit binary, feed that to a D/A converter IC, then you can have your 18 voltage steps. Then make a single output stage with a high voltage OP amp and an output transistor, where its output is fed back to the OP amp -input to cancel the temperature dependency of the output device, so as to scale up the voltages. Then, if you really do need 18 separate physical voltage lines , you can just use 18 diodes in an OR configuration and place one of the same sort of diodes in the OP amp circuit to cancel the temperature coefficient of the output diodes. Looking at the Bolgspot circuit it seems that you just need 18 levels applied to a single line, but obviously the diodes would be needed if you want selection by the physical switch too.

Martin G7MRV 22nd Jan 2019 11:11 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
The potentiometer chain the voltages are selected from is part of the original equipment, I cant replace that as it also comprises the temperature compensation components of the radios tuning system. I just have to replace the 18way switch that feeds the selected voltage to the receivers varicaps. Im not starting with digital levels as such, but the control circuit im building will produce 3v logic levels to drive the switching.

Each of the 18 voltages sets the varicap tuning for one of the 10MHz ranges the receiver covers.

Argus25 22nd Jan 2019 11:32 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Ok, that makes it clear now what you are doing.

It is important then not to draw any significant current from the divider chain.

One way that would work, which solves all the problems but one, would be to use miniature reed relays, the coils being easily controlled by the logic circuit and would give all the isolation required, but I suspect there will not be room. Using the mosfet and driver transistor, you would also need a fairly high value of gate source tie off resistor, so the load on the line you were switching was very low.

Martin G7MRV 22nd Jan 2019 11:44 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Relays were my first thought, but as you say take up considerable space. But then again, it looks like im looking at 36 transistors plus discretes, which isnt going to be small either! Ive also pondered opto-isolators, as there are some very small ones available.

One thing I should have said, as it might have a bearing - its not my intention to use the actual divider chain from the radio, as those resistors are mounted on the switch wafer, but to switch the connections to either end of it, wo a replica divider chain, and select my voltage taps from that (it cuts the wiring down dramatically!) So, I already have a 24v 4PDT relay involved taking those connections, and the connection from the switch moving contact, with the 4th pole switching out the 0v connection to the synthesiser control switches to allow my logic to take over.

Perhaps suitable SMT relays might allow me to fit enough in?

The radio is a Clansman PRC344 UHF AM manpack. At present, the only available space comes from where two plug in modules have been removed to disable the transmitter (sensible as it cant be used at all), giving me maybe a few square inches. Which is why some form of IC switch would be ideal!

Martin G7MRV 22nd Jan 2019 11:57 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Actually, now I think about it, perhaps the opto-couplers are the way to go? Easy to drive from logic, good isolation, small package size, and amazingly cheap. Just seen 20x PC817s on the slow boat from the orient for a quid, 80v Vce. Might be right on the limit for the top voltage selection, but the current through them would be minimal.

These are bipolar of course, but would the junction voltage drop make any real difference in this case? The ends of the divider chain go to presets used to adjust the highest and lowest voltages, perhaps I could just make a slight adjustment that had minimal effect on the filter tuning.

I could probably shoehorn 18 optoisolators into the space left by one of the removed modules, plus the logic chips and the PIC uP, leaving the other modules space for the relay and power supply...

(found 100x PC817 optocouplers for under £2! so ive bought them anyway!)

Argus25 23rd Jan 2019 12:12 am

Re: High side high voltage MOSFET switching with 3V logic drive?
 
In a project a while back I used some ASSR-1228-002E "solid state relays" that have a photovoltaic diode coupled to output mosfets:

https://www.broadcom.com/products/op...assr-1228-002e

Looking at these again, the voltage rating for the "switch" is 60V, but there might be higher voltage versions. If you used these there is no switch voltage drop to worry about and the isolation is excellent.

They can be seen in this project on page 8:

http://worldphaco.com/uploads/INCREA...OR_TO_AUTO.pdf

It looks like they have 400V ones that would be just fine, they have a low on resistance:

https://www.broadcom.com/products/op...assr-4120-002e

Martin G7MRV 23rd Jan 2019 12:30 am

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Those do look ideal, but a quick search shows them to be about two quid a throw! Would make for a very expensive project. Im only really doing it as a proof of concept (because someone challenged me to make a vintage military radio scan!!!)

I'll try the bipolar optocouplers, and see what affect they have on the performance of the receiver, and whether any effect can be trimmed out. I could probably rig just one coupler up on the radio to see what occurs, before building the rest of the project. If it needs the MOSFET types i'll have to reconsider whether its worth the cost.

Guest 23rd Jan 2019 3:40 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
0..3V in and an amplifier?

G6Tanuki 23rd Jan 2019 3:50 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
In this instance I'd be using relays - rather than reeds I'd use the type that come in a TO5 transistor-style can. I used these extensively for RF switching in something I designed a few decades ago.

Teledyne is the name to search for. They do latching versions which are rather handy in circumstances where you have to rely on battery power.

SiriusHardware 23rd Jan 2019 4:49 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

Originally Posted by Martin G7MRV (Post 1113414)
These are bipolar of course, but would the junction voltage drop make any real difference in this case?

As long as you drive the optocouplers fully on, there should be next to zero voltage drop across the collector-emitter terminals. You're maybe thinking of the base to emitter forward voltage which doesn't apply here. Obviously, they are polarity sensitive devices so the collector of the output transistor has to go to the 'most positive' side of the gap which the transistor will be bridging.

The only voltage drop to worry about here is the forward voltage drop of the photodiode, where you will need more than a specific minimum voltage to turn the diode on. I generally get away with using 330R series resistors when driving optocoupler inputs from 3V microprocessor pins.

Martin G7MRV 23rd Jan 2019 5:35 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Yes your right, im thinking of Vbe.

The photodiode in the PC817 states Vf of 1.2-1.4v, at a test If of 5mA, max 20mA, which does indeed come out at 330R for 3v 5mA, and as im only activating one out of the 18 at a time, i'll be able to get away with just the one resistor.

My remaining worry then for this part of the project (apart from whether I can physically fit the whole lot into the radio) is whether the 80v maximum will be exceeded or approached too closely on any of the higher ranges. I'll need to actually measure the voltages out of each step of the divider chain to establish that, but if so it will just mean I have to choose a higher rated device for those lines

Guest 23rd Jan 2019 7:00 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

but if so it will just mean I have to choose a higher rated device for those lines
or two (if bulk buy makes them cheap) in series.

Martin G7MRV 23rd Jan 2019 7:05 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Hmm, never thought of putting optocouplers in series...

My bulk buy is £1.95 for 100x PC817, so certainly cheap enough! (although at that price i'll be running them all through a test jig!)

It would mean that the series pair would need their own series resistor for the LED, but I think I can live with that!

Guest 23rd Jan 2019 7:48 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

although at that price i'll be running them all through a test jig
Would it be worth testing a few at, say, 150V or until destruction to see the real voltage capability?

Guest 23rd Jan 2019 7:50 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Just read the datasheet for the PC817, it says 80V so all should be well.

Martin G7MRV 23rd Jan 2019 8:36 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Yes 80v max. The divider chain runs from 80v, but i'll have to check how close to that the highest tap on the chain is. I suspect it will be 'just' on the limit,


All this because of a late night drunken challenge from a mate - "I could make that scan...", "Go on then!"....

Martin G7MRV 25th Jan 2019 10:43 am

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

Originally Posted by Martin G7MRV (Post 1113684)
Yes 80v max. The divider chain runs from 80v, but i'll have to check how close to that the highest tap on the chain is. I suspect it will be 'just' on the limit,

Measured the tuning HT rail this morning, highest it reaches is a tad over 65v, so well within the rating of the PC817

Martin G7MRV 25th Jan 2019 6:52 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Quote:

Originally Posted by Argus25 (Post 1113391)

If you are starting with digital signals, one way to do this would be to encode that into 5 bit binary, feed that to a D/A converter IC, then you can have your 18 voltage steps. Then make a single output stage with a high voltage OP amp and an output transistor, where its output is fed back to the OP amp -input to cancel the temperature dependency of the output device, so as to scale up the voltages. Then, if you really do need 18 separate physical voltage lines , you can just use 18 diodes in an OR configuration and place one of the same sort of diodes in the OP amp circuit to cancel the temperature coefficient of the output diodes. Looking at the Blogspot circuit it seems that you just need 18 levels applied to a single line, but obviously the diodes would be needed if you want selection by the physical switch too.

Ive been thinking about this idea. The radio in question is early 1980s vintage, so im wondering if the internal temperature compensation used to stabilise the divider chain, really needs to be retained? So long as whatever I replace the divider chain with is also reasonably temperature stable. Ultimately im going to be controlling this selection with a PIC microcontroller, a 16F628A, which can provide PWM outputs. Ive seen circuits that use this PWM to feed an RC filter and an op-amp to drive an LM317 adjustable regulator, which if fed with its max input 37v can give a variable voltage of about 1 - 35v, depending on the PWM duty cycle,

My only problem with this approach is the need to go up to close to 70v with an 80v source (actual max measured voltage is 65v), so the LM317 isnt suitable, but is there another device that can replace it? The current is minimal, so would it work using the op-amp to drive a transistor?

This could potentially simplify the build drastically!

Thanks for the heads up on this technique - im going to look into it now!

Guest 25th Jan 2019 7:26 pm

Re: High side high voltage MOSFET switching with 3V logic drive?
 
Why not do the PWM at 80V and then smooth? If the tuning current is low then a simple MOSFET/resistor amplifier would do the 80V bit and then a two/three stage low pass do the smoothing.


All times are GMT +1. The time now is 2:21 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.