[Marcel Baum]

I have got a linear pump which needs a sinusoidal excitation.
The pumpflow is controlled by the amplitude of the AC to the coils
The amplitude is controlled via a DC input signal which on itself can change
very fast.
So far I have used a 24VAC pump, which is no longer available, the actual
pumps have 230VAC (50hz). The rated power is roughly 150W.

My present design is powered from 24V DC and utilises a full (H-) Mosfet
bridge.
The bridge is controlled by a MC 33035 which is a PWM motor controller with
an analogue input and a direction signal. The analogue input gets a sine
have wave with variable amplitude, the DIR input sees a 50hz TTL signal in
phase with the sinus.
To go on with this design I would have to rectify the 230V, change to a 500V
Mosfet bridge and to add some HV photocouplers to feed the hi side Fets of
the bridge.

Does anyone have better suggestion how to directly modulate the mains AC
instead of synthesising the AC from a DC source?
I probably should mention that the pump wants to see a sinus like (or
trapezoid) excitation which limits the use of a triac .

Any ideas are welcome.

 

[Joerg]

I have got a linear pump which needs a sinusoidal excitation.
The pumpflow is controlled by the amplitude of the AC to the coils
The amplitude is controlled via a DC input signal which on itself can change
very fast.
So far I have used a 24VAC pump, which is no longer available, the actual
pumps have 230VAC (50hz). The rated power is roughly 150W.

My present design is powered from 24V DC and utilises a full (H-) Mosfet
bridge.
The bridge is controlled by a MC 33035 which is a PWM motor controller with
an analogue input and a direction signal. The analogue input gets a sine
have wave with variable amplitude, the DIR input sees a 50hz TTL signal in
phase with the sinus.
To go on with this design I would have to rectify the 230V, change to a 500V
Mosfet bridge and to add some HV photocouplers to feed the hi side Fets of
the bridge.

That's pretty much how it's done. But optocouplers are often not used.
One common method is to generate a "helper voltage" slightly above each
rail, then use all n-channel FETs (cost reasons) and drive these with
BJTs (also cost reasons). Ok, we have it easier over here because our
line voltage is 120V so we can get away with MSPA42 et cetera which can
be had for around 3 Cents, plus whatever FET is on sale. It'll cost more
over there if you must use a 230V motor.

Does anyone have better suggestion how to directly modulate the mains AC
instead of synthesising the AC from a DC source?

Not one that would be cheaper.

I probably should mention that the pump wants to see a sinus like (or
trapezoid) excitation which limits the use of a triac .

Any ideas are welcome.

Oh, another cost savings hint: The KA3525 costs less than the MC33035. A
whole lot less.

 

[Joerg]

I have got a linear pump which needs a sinusoidal excitation.
The pumpflow is controlled by the amplitude of the AC to the coils
The amplitude is controlled via a DC input signal which on itself can change
very fast.
So far I have used a 24VAC pump, which is no longer available, the actual
pumps have 230VAC (50hz). The rated power is roughly 150W.

My present design is powered from 24V DC and utilises a full (H-) Mosfet
bridge.
The bridge is controlled by a MC 33035 which is a PWM motor controller with
an analogue input and a direction signal. The analogue input gets a sine
have wave with variable amplitude, the DIR input sees a 50hz TTL signal in
phase with the sinus.
To go on with this design I would have to rectify the 230V, change to a 500V
Mosfet bridge and to add some HV photocouplers to feed the hi side Fets of
the bridge.

Does anyone have better suggestion how to directly modulate the mains AC
instead of synthesising the AC from a DC source?
I probably should mention that the pump wants to see a sinus like (or
trapezoid) excitation which limits the use of a triac .

Any ideas are welcome.

Just to avoid a misunderstanding: The PWM is usually not done with the
bridge itself. It is done by a converter up front and that gets the
sine-wave control signal. The H-bridge just switches.

 

[Marcel Baum]

Just to avoid a misunderstanding: The PWM is usually not done with the
bridge itself. It is done by a converter up front and that gets the
sine-wave control signal. The H-bridge just switches.

Thank you for your response

The 33035 has got a pwm on chip, the high side Fet is responsible for the
direction the corresponding low side one is pulse modulated according to
the actual input voltage.
Its not that expensive here.
Its very much like in the LMD 18245 used with stepper motors (this one is
really expensive).

What I was thinking about is to use an existing DC-AC converter (e.g. 24VDC
to 230VAC) lay out and to add a control input for the output voltage. No
idea if this exists already. Have you ever looked into something similar?

Since this pump is part of a medical life supporting device it has to run
on batteries for at least 20 minutes anyway. The most simple way is to run
it from a UPS (having its internal batteries), or if not possible use a
buffered 24V battery with a DC-AC step up converter.

 

[Joerg]

Thank you for your response

The 33035 has got a pwm on chip, the high side Fet is responsible for the
direction the corresponding low side one is pulse modulated according to
the actual input voltage.
Its not that expensive here.
Its very much like in the LMD 18245 used with stepper motors (this one is
really expensive).

You want to pulse the bridge to get the sine wave and rely on the motor
windings to smooth things out? That can backfire the minute the motor
manufacturer changes the design. Personally I wouldn't do that. EMC can
also become quite a nightmare, as can motor noise.

What I was thinking about is to use an existing DC-AC converter (e.g. 24VDC
to 230VAC) lay out and to add a control input for the output voltage. No
idea if this exists already. Have you ever looked into something similar?

Well, they work pretty much like I described. A switcher around a 3525
generates a voltage between zero and the desired peak voltage, usually
as a forward converter. Sinusoidal half-waves. Then the bridge toggles that.

Since this pump is part of a medical life supporting device it has to run
on batteries for at least 20 minutes anyway. The most simple way is to run
it from a UPS (having its internal batteries), or if not possible use a
buffered 24V battery with a DC-AC step up converter.

I design medical devices for a living and running an AC motor at over
50V is usually frowned upon by the agency guys. Are you sure you can't
get a 24V version anywhere? Also, nearly none of the parts vendors
endorses life support applications but I am pretty sure you know that.

 

[john jardine]

I have got a linear pump which needs a sinusoidal excitation.
The pumpflow is controlled by the amplitude of the AC to the coils
The amplitude is controlled via a DC input signal which on itself can change
very fast.
So far I have used a 24VAC pump, which is no longer available, the actual
pumps have 230VAC (50hz). The rated power is roughly 150W.

My present design is powered from 24V DC and utilises a full (H-) Mosfet
bridge.
The bridge is controlled by a MC 33035 which is a PWM motor controller with
an analogue input and a direction signal. The analogue input gets a sine
have wave with variable amplitude, the DIR input sees a 50hz TTL signal in
phase with the sinus.
To go on with this design I would have to rectify the 230V, change to a 500V
Mosfet bridge and to add some HV photocouplers to feed the hi side Fets of
the bridge.

Does anyone have better suggestion how to directly modulate the mains AC
instead of synthesising the AC from a DC source?
I probably should mention that the pump wants to see a sinus like (or
trapezoid) excitation which limits the use of a triac .

Any ideas are welcome.

Google for "AC Chopper". There's 2 very useful papers out there (I can't be
arsed looking). They handle the AC directly.
ST also have a useful application note

 

[Marcel Baum]

You want to pulse the bridge to get the sine wave and rely on the motor
windings to smooth things out? That can backfire the minute the motor
manufacturer changes the design. Personally I wouldn't do that. EMC can
also become quite a nightmare, as can motor noise.

Its not a common motor. Its a pair of coils with an oscillating permanent
magnet assembly in between, which actuates two membrane heads.
I use this circuit since 10 years without any problems.

Well, they work pretty much like I described. A switcher around a 3525
generates a voltage between zero and the desired peak voltage, usually as
a forward converter. Sinusoidal half-waves. Then the bridge toggles that.



I design medical devices for a living and running an AC motor at over 50V
is usually frowned upon by the agency guys. Are you sure you can't get a
24V version anywhere? Also, nearly none of the parts vendors endorses life
support applications but I am pretty sure you know that.

The pumps are not close to the patient and they are mounted totally floating
in separate compartments, what counts is their leakage current to ground.
Since I have an UPS in front, this will mainly determine the leakage
current.
I had 24V pumps so far, but all the far east manufacturers of this pump
principle stopped the production for this supply voltage. Manufacturers in
far east are not that restrictive when applying their products in LSS.

I have 3 pumps in my system, it would be an awful lot of work to strip them
down to a level at which I can remove the coils, moreover I would loose the
manufacturers certificate for the use in an oxygen enriched environment as
soon as I open the hood.
Furthermore at this stage my system only goes into veterinary medicine, I
have not got the infrastructure to manufacture human medicine devices
myself.

 

[Marcel Baum]

Google for "AC Chopper". There's 2 very useful papers out there (I can't
be
arsed looking). They handle the AC directly.
ST also have a useful application note

John this was really a good hint.

Not only that ST has an application note , they also have a ready to go
evaluation board which perfectly fits my application (STEVAL-IHM006V1).

I just will have to change the potentiometer to an analogue input signal and
it should run perfectly. Its even not so expensive, its about 50 Euro.

Thanks a lot for giving me this hint.

Marcel

 

[Joerg]

Its not a common motor. Its a pair of coils with an oscillating permanent
magnet assembly in between, which actuates two membrane heads.
I use this circuit since 10 years without any problems.

Ah, like an electric razor "motor"? Then you should be able to even
shield the coils if necessary.

The pumps are not close to the patient and they are mounted totally floating
in separate compartments, what counts is their leakage current to ground.
Since I have an UPS in front, this will mainly determine the leakage
current.
I had 24V pumps so far, but all the far east manufacturers of this pump
principle stopped the production for this supply voltage. Manufacturers in
far east are not that restrictive when applying their products in LSS.

I have 3 pumps in my system, it would be an awful lot of work to strip them
down to a level at which I can remove the coils, moreover I would loose the
manufacturers certificate for the use in an oxygen enriched environment as
soon as I open the hood.

Yes, loss of that certificate would certainly be a bummer.

Furthermore at this stage my system only goes into veterinary medicine, I
have not got the infrastructure to manufacture human medicine devices
myself.

When first designing something for veterinary use I was surprised that
the rules were pretty much as strict as for normal med use, at least in
the US. Personally I regard the life of an animal as much as that of a
human. The laws may be different sometimes but when they used one of our
ultrasound machines on a million Dollar race horse I realized what can
be at stake.

 

[Jan Panteltje]

On a sunny day (Wed, 24 Oct 2007 14:49:54 GMT) it happened Joerg

human.

mm so you are a vegetarian, and let mosquitos sting you, sticking your
arm out to feed them?

 

[Joerg]

On a sunny day (Wed, 24 Oct 2007 14:49:54 GMT) it happened Joerg


mm so you are a vegetarian, and let mosquitos sting you, sticking your
arm out to feed them?

Nope, doesn't go that far

 

[Jan Panteltje]

Nope, doesn't go that far

OK, now I know where you stand.

 

[Marcel Baum]

On a sunny day (Wed, 24 Oct 2007 14:49:54 GMT) it happened Joerg


mm so you are a vegetarian, and let mosquitos sting you, sticking your
arm out to feed them?

Not necessary at all. In my case, they will find their way without any hint.

 

[Marcel Baum]

Ah, like an electric razor "motor"? Then you should be able to even shield
the coils if necessary.

Exactly, and the entire drive is capsuled in an alloy cast, which again sits
in a larger tight cabinet. And in addition this pump stands in a closed
iron box, mainly for noise and safety reasons. If there is any EMI it
wouldn't come out of this 3 fold cage.

Yes, loss of that certificate would certainly be a bummer.

It is hard enough to persuade them to deliver it with a written certificate.
Oxygen is very special, there is a standard defining the maximum amount of
grease (oil, fat) per cm^2 of surface in contact with oxygen. Furthermore
the material for your membranes , gaskets and sealings have to comply with
the standards for oxygen.

When first designing something for veterinary use I was surprised that the
rules were pretty much as strict as for normal med use, at least in the
US. Personally I regard the life of an animal as much as that of a human.
The laws may be different sometimes but when they used one of our
ultrasound machines on a million Dollar race horse I realized what can be
at stake.

Interestingly we have no standards for veterinary medical devices, beside
the general IEC 601-1. The requirements for breathing machines used in human
anaesthesia are not obligatory for animal use. Anyway my machine will
fulfil all this requirements, however the effort in testing the software is
by far easier compared to human applications, as is the entire
documentation.
Today a one million Dollar horse is put on a horse anaesthesia machine,
which is roughly 40 years behind its equivalent for human use. That's one of
the reasons why a huge number of horses are experiencing lung problems after
a operation of 2 hours.
But this I guess is far off topic.

 

[Joerg]

Exactly, and the entire drive is capsuled in an alloy cast, which again sits
in a larger tight cabinet. And in addition this pump stands in a closed
iron box, mainly for noise and safety reasons. If there is any EMI it
wouldn't come out of this 3 fold cage.




It is hard enough to persuade them to deliver it with a written certificate.
Oxygen is very special, there is a standard defining the maximum amount of
grease (oil, fat) per cm^2 of surface in contact with oxygen. Furthermore
the material for your membranes , gaskets and sealings have to comply with
the standards for oxygen.



Interestingly we have no standards for veterinary medical devices, beside
the general IEC 601-1. The requirements for breathing machines used in human
anaesthesia are not obligatory for animal use. Anyway my machine will
fulfil all this requirements, however the effort in testing the software is
by far easier compared to human applications, as is the entire
documentation.

Well, you guys don't have to deal with the FDA. That's why us guys often
start clinical trials in Europe

Today a one million Dollar horse is put on a horse anaesthesia machine,
which is roughly 40 years behind its equivalent for human use. That's one of
the reasons why a huge number of horses are experiencing lung problems after
a operation of 2 hours.
But this I guess is far off topic.

That's very different here. We have one of the larger vet clinics at UC
Davis right here in town. They are using the latest and greatest gear on
horses and other animals, not 40 year old stuff. Don't know about
anesthesia though, haven't looked at those machines but nothing in that
lab looked older than five years.

 

[Marcel Baum]

Well, you guys don't have to deal with the FDA. That's why us guys often
start clinical trials in Europe

I wouldn't say that the regulations here are by any means easier.
FDA is much more looking for procedures, whereas here functionality and
first fault safety is a much stronger issue.
Clinical trials now can only be started with a device having already a
complete CE mark documentation. But this is only true for well established
methods. Once you plan to introduce a new therapy mode which is different
from state of the art, you need to go through extensive animal experiments
(and later on multicentre patient studies) to demonstrate its benefits and
its possible side effects.
That's why the number of manufacturers in this field was massively shrinking
during the last years. Documentation and quality control is up to about 40%
of the total development costs and unaffordable for a smaller company.

That's very different here. We have one of the larger vet clinics at UC
Davis right here in town. They are using the latest and greatest gear on
horses and other animals, not 40 year old stuff. Don't know about
anesthesia though, haven't looked at those machines but nothing in that
lab looked older than five years.

This was not my point. Even the most recent anaesthesia machines deliver a
therapy quality comparable to what I found in human medicine, when I stepped
into this field some 38 years ago.
All the developments resulting from the better understanding of lung
physiology never were transferred into large animal anaesthesia machines. I
have spent 28 years in this research field, and I now will try to introduce
this experience into veterinary medicine.

 

[Joerg]

Marcel Baum wrote:

[...]

This was not my point. Even the most recent anaesthesia machines deliver a
therapy quality comparable to what I found in human medicine, when I stepped
into this field some 38 years ago.
All the developments resulting from the better understanding of lung
physiology never were transferred into large animal anaesthesia machines. I
have spent 28 years in this research field, and I now will try to introduce
this experience into veterinary medicine.

Well, with that much experience you should be able to get it done.
Compared to you I probably count as a newbie, in medical devices I've
only got 21 years under the belt ;-)