Adapting An Electrophoresis HVDC Power Supply To Microampere Level

[SMH]

I had posted this problem earlier, but without detailing much how the power
supply is to be used.

I have written a perhaps more detailed web page that explains how it is to
be used.

I put this question to a professor of electrical engineering who is an
expert in circuit design, but he could not give much time to it and wrote a
schematic that was short of what I needed, such as which parts to buy (name
and/or part number) so that I can build something to test. I put his
explanation and drawn schematic on this web page. If someone can actually
name the part numbers, I can buy and put them together for testing (if you
can suggest where I buy them too, that would be helpful).

http://tinyurl.com/3yxzev

I was looking for a low cost (say no more than a couple of hundred
dollars), easily implementable solution.

That solution could be buying an already assembled set of components in the
box from a maker, or building something on a circuit (prototype?) board
with components piece-by-piece...opamp, MOSFET, resistor, capacitor, etc.

It must be a solution that requires that the person has only a fundamental
understanding of the physics of electricity, and who may know how to use a
soldering iron.

 

[linnix]

I had posted this problem earlier, but without detailing much how the power
supply is to be used.

I have written a perhaps more detailed web page that explains how it is to
be used.

I put this question to a professor of electrical engineering who is an
expert in circuit design, but he could not give much time to it and wrote a
schematic that was short of what I needed, such as which parts to buy (name
and/or part number) so that I can build something to test. I put his
explanation and drawn schematic on this web page. If someone can actually
name the part numbers, I can buy and put them together for testing (if you
can suggest where I buy them too, that would be helpful).

http://tinyurl.com/3yxzev

I was looking for a low cost (say no more than a couple of hundred
dollars), easily implementable solution.

That solution could be buying an already assembled set of components in the
box from a maker, or building something on a circuit (prototype?) board
with components piece-by-piece...opamp, MOSFET, resistor, capacitor, etc.

It must be a solution that requires that the person has only a fundamental
understanding of the physics of electricity, and who may know how to use a
soldering iron.

Interesting. I have not seen too many semiconductors with more than
1KV breakdown, including MOSFET. You will likely need a chains of
MOSFET in series just to spread out the Drain Source breakdown.
Setting the proper gate control voltage is another problem. To echo
another thread: Relays?

 

[Robert Baer]

I had posted this problem earlier, but without detailing much how the power
supply is to be used.

I have written a perhaps more detailed web page that explains how it is to
be used.

I put this question to a professor of electrical engineering who is an
expert in circuit design, but he could not give much time to it and wrote a
schematic that was short of what I needed, such as which parts to buy (name
and/or part number) so that I can build something to test. I put his
explanation and drawn schematic on this web page. If someone can actually
name the part numbers, I can buy and put them together for testing (if you
can suggest where I buy them too, that would be helpful).

http://tinyurl.com/3yxzev

I was looking for a low cost (say no more than a couple of hundred
dollars), easily implementable solution.

That solution could be buying an already assembled set of components in the
box from a maker, or building something on a circuit (prototype?) board
with components piece-by-piece...opamp, MOSFET, resistor, capacitor, etc.

It must be a solution that requires that the person has only a fundamental
understanding of the physics of electricity, and who may know how to use a
soldering iron.

In theory, that circuit will work, except (as previously noted) the
FET will not be able to withstand more than 1000V.
Granted that a number of FETs can be daisy-chained (or cascoded), but
that can be risky.
Why not start with an AC to DC supply where the level of the DC is
continuously adjustable from zero to full value by adjusting the AC input?
Now a FWB in series with the AC input can allow the DC equivalent of
a variable resistor to be used for that adjusting, and the maximum
voltage (120VAC line) could allow a FET with a 400V rating act as that
variable resistor.
Does that sound like a step in the right direction?
If so, someone here could design and make a number of these beasties...

 

[linnix]

In theory, that circuit will work, except (as previously noted) the
FET will not be able to withstand more than 1000V.
Granted that a number of FETs can be daisy-chained (or cascoded), but
that can be risky.
Why not start with an AC to DC supply where the level of the DC is
continuously adjustable from zero to full value by adjusting the AC input?
Now a FWB in series with the AC input can allow the DC equivalent of
a variable resistor to be used for that adjusting, and the maximum
voltage (120VAC line) could allow a FET with a 400V rating act as that
variable resistor.
Does that sound like a step in the right direction?
If so, someone here could design and make a number of these beasties...

Yes, it is in the right direction. I would use a micro based DC-DC
booster (the oscillator side is really AC anyway), since the OP is
going to need a voltage sequencer from 500V to 10KV. He can use op-
amp current sensing and some resistors (0.1% or better) voltage
sensing. Basically, the OP is asking for a programmable high voltage
power supply.

 

[SMH]

Yes, it is in the right direction. I would use a micro based DC-DC
booster (the oscillator side is really AC anyway), since the OP is
going to need a voltage sequencer from 500V to 10KV. He can use op-
amp current sensing and some resistors (0.1% or better) voltage
sensing. Basically, the OP is asking for a programmable high voltage
power supply.

A fully programmable power supply would include the ability to set several
steps at different voltages (including ramping up and down as well as
constant voltages) for certain durations (i.e., setting volt-hours).

I was just trying to build in some intelligence into the circuit that
would provide safety, i.e. a shutdown of the voltage to prevent burning up
of the gel strip and potentially a lab fire.

As far as components dependent upon AC, I had not planned on making any AC
available to the circuit: it was going to plug in strictly to the HVDC
power supply and use that, branching off for any control or regulating
circuits it might need.

I realize now that the thing I am asking for is much more complex than I
imagined. The goal of this project was effectively to try to make use of
lots of electrophoresis power supplies out there that can be used for
proteomics and genomics where electrophoresis is done at smaller
(microampere vs. milliampere scales); think of it as a "recycling
project." But now these power supplies will be put on the junk heap as
they find more limited application.

There are already commercial power supplies out there that provide
microampere-monitoring HVDC, and basically two vendors sell their units
for about $8000-12,000. I was hoping that a conversion/modfication could
be done for 1/50 of the cost.

I tried to get some EE grad students interested in the project (posted to
mailing list), thinking it would take them minutes to solve the problem
and they would get a publication out of it; one response from someone who
wanted to know more and I haven't heard back. Of course, if the effort
requires more time, students like these would want monetary support, which
I am unable to offer (else if I had that much money, I would just buy the
commercial product). If there is a specific workable solution to this, it
clearly requires me to take a couple of semesters of circuit analysis and
design at the university, and my deadline is pretty much the middle of
this month (passed, that is). At this point, I have concluded it will
have to be someone else's work that pushes forward this idea since I have
run out of time will no longer have the means to move it forward (my
access to the equipment comes to an end at the end of the month).

 

[Winfield Hill]

A fully programmable power supply would include the ability to set several
steps at different voltages (including ramping up and down as well as
constant voltages) for certain durations (i.e., setting volt-hours).

I was just trying to build in some intelligence into the circuit that
would provide safety, i.e. a shutdown of the voltage to prevent burning up
of the gel strip and potentially a lab fire.

As far as components dependent upon AC, I had not planned on making any AC
available to the circuit: it was going to plug in strictly to the HVDC
power supply and use that, branching off for any control or regulating
circuits it might need.

I realize now that the thing I am asking for is much more complex than I
imagined. The goal of this project was effectively to try to make use of
lots of electrophoresis power supplies out there that can be used for
proteomics and genomics where electrophoresis is done at smaller
(microampere vs. milliampere scales); think of it as a "recycling
project." But now these power supplies will be put on the junk heap as
they find more limited application.

There are already commercial power supplies out there that provide
microampere-monitoring HVDC, and basically two vendors sell their units
for about $8000-12,000. I was hoping that a conversion/modfication could
be done for 1/50 of the cost.

I tried to get some EE grad students interested in the project (posted to
mailing list), thinking it would take them minutes to solve the problem
and they would get a publication out of it; one response from someone who
wanted to know more and I haven't heard back. Of course, if the effort
requires more time, students like these would want monetary support, which
I am unable to offer (else if I had that much money, I would just buy the
commercial product). If there is a specific workable solution to this, it
clearly requires me to take a couple of semesters of circuit analysis and
design at the university, and my deadline is pretty much the middle of
this month (passed, that is). At this point, I have concluded it will
have to be someone else's work that pushes forward this idea since I have
run out of time will no longer have the means to move it forward (my
access to the equipment comes to an end at the end of the month).

I have no idea why the commercial low-current HV supplies
are so expensive. Unless it's liability insurance?

As for recycling those old HV supplies, unless there's
something special about their control and display systems,
I'd say it'd be cheaper and easier to start from scratch.

While it's difficult to make a 5 to 10 or 15kV regulator
from series MOSFETs, it's easy to make an feedback power
controller incorporating a HV dc-dc converter. Usually
I simply use HV resistors, etc., with feedback opamps
driving the ref input of a common LM317 three-terminal
regulator IC, which in turn drives the dc-dc converter's
power input, e.g. 12V for full-output voltage, etc.

For low-current rather than voltage regulation, provide
the machine's feedback from a current sense resistor on
the low side of the cheap DC-DC converter module, and
add a high-voltage ballast resistor on the high side.

I'd also add safety shutoff stuff.

As far as access to the old equipment is concerned, this
is an issue only if there's truly something valuable there.
Something other than the HV generator portion, that is.
Not unless it can be controlled in a fashion similar to
the easy approach with dc-dc converter modules.

 

[GregS]

I have no idea why the commercial low-current HV supplies
are so expensive. Unless it's liability insurance?

As for recycling those old HV supplies, unless there's
something special about their control and display systems,
I'd say it'd be cheaper and easier to start from scratch.

While it's difficult to make a 5 to 10 or 15kV regulator
from series MOSFETs, it's easy to make an feedback power
controller incorporating a HV dc-dc converter. Usually
I simply use HV resistors, etc., with feedback opamps
driving the ref input of a common LM317 three-terminal
regulator IC, which in turn drives the dc-dc converter's
power input, e.g. 12V for full-output voltage, etc.

For low-current rather than voltage regulation, provide
the machine's feedback from a current sense resistor on
the low side of the cheap DC-DC converter module, and
add a high-voltage ballast resistor on the high side.

I'd also add safety shutoff stuff.

As far as access to the old equipment is concerned, this
is an issue only if there's truly something valuable there.
Something other than the HV generator portion, that is.
Not unless it can be controlled in a fashion similar to
the easy approach with dc-dc converter modules.

The power supplies I mostly see are the older high current type, which
some start out just below $1000. I have not worked on or seen any
of the ultra low current types. I imagine noise and current sensing
are more difficult to control and monitor, but should be within reason.
I just got done fixing the analog meter movement in a 500 volt 500 ma.
EC Apparatus unit.

There is always someone around here who wants to do stuff and they say they
have no money.

greg


greg