Hi everyone,
This is my first post here and its not an audio amp but this still seemed like a good spot. I've been doing alot of reading from here thinking about designs.
I am working on designing a push pull AB piezo amplifier. The piezo needs to see from ~ -30 to 150 volts and I would like to be able to output around 100mA (although futures designs may call for more maybe only 150 volts though). Noise is extremely important however the output will be filtered low ~300 Hz. I'm pretty new at discrete component designs however I am kind of familiar with bjts. There doesn't seem to be alot that will meet my power requirements (PNP at least) I was looking at MJ21195 but where I want to operate puts me pretty close to the edge of the SOA. I can clamp the voltage so it won't exceed specs and I don't mind putting a heavy duty heatsink to keep it cool.
Do you think this would be acceptable? Or should I be looking into other technologies. I've heard mosfets are noisy but have no real data. I've heard of higher voltage jfets but are they readily available? No real P channel IGBTs so that seems like it would complicate my design. I've also heard there not good for linear amps for various reason but have not worked with them much.
Any thoughts are appreciated.
Thanks
Russell
This is my first post here and its not an audio amp but this still seemed like a good spot. I've been doing alot of reading from here thinking about designs.
I am working on designing a push pull AB piezo amplifier. The piezo needs to see from ~ -30 to 150 volts and I would like to be able to output around 100mA (although futures designs may call for more maybe only 150 volts though). Noise is extremely important however the output will be filtered low ~300 Hz. I'm pretty new at discrete component designs however I am kind of familiar with bjts. There doesn't seem to be alot that will meet my power requirements (PNP at least) I was looking at MJ21195 but where I want to operate puts me pretty close to the edge of the SOA. I can clamp the voltage so it won't exceed specs and I don't mind putting a heavy duty heatsink to keep it cool.
Do you think this would be acceptable? Or should I be looking into other technologies. I've heard mosfets are noisy but have no real data. I've heard of higher voltage jfets but are they readily available? No real P channel IGBTs so that seems like it would complicate my design. I've also heard there not good for linear amps for various reason but have not worked with them much.
Any thoughts are appreciated.
Thanks
Russell
FETs noisy? Maybe, but if you use a more-or-less standard audio amp topology, the noise will be dominated by the input stage. Does this amp need to be run open-loop? If global feedback is used I don't see any reason you couldn't use MOSFETs - even an all-N-channel design.
If you did want to stay away from FETs, and are unsure if the s/b limited SOA of even the most rugged bipolars is good enough, you could use a series connected output stage.
If you did want to stay away from FETs, and are unsure if the s/b limited SOA of even the most rugged bipolars is good enough, you could use a series connected output stage.
Apex PA08A
PA08 - 300V, 150mA, Low Bias Power Amplifier
I have a bunch of these if you want to buy one. Still needs a power supply but it may fit your bill.
PA08 - 300V, 150mA, Low Bias Power Amplifier
I have a bunch of these if you want to buy one. Still needs a power supply but it may fit your bill.
I need to design my own amp.
The amp will be closed loop and have a gain of around 15 something to get my input signal +-12ish to the 150V I need. The drive is offset -30 to 150 volts I think thats what youre asking
The piezo is a stacked actuator, the ones I use are 1.8uF they may get 2-3x larger. 200 ohm series resistor for the filter sometimes larger.
I'm not an expert but I was planning on using an opamp input, a gain stage(common emiter) maybe something else and then into a bias and output scheme and have the output feedback to the opamp and let it do the work. I could post up a circuit in a day or so I have been doing some simulations with bjts outputs but decided to step back when I realized there weren't any that could really handle what I need.
Heres a packaged amp I use PA15FL the price doubled when they were bought out by cirrus logic. The standard deviation noise brought back in is about 2ppm filtered at 200 Hz so this needs to be close (its in a closed loop system).
I will be sending out some pcbs next weeks so if I get something decent put together in the simulation I will panel it and not have to prototype my first trial
Again all thoughts are welcome I have no real experience designing these things just been reading up for the last couple months.
The amp will be closed loop and have a gain of around 15 something to get my input signal +-12ish to the 150V I need. The drive is offset -30 to 150 volts I think thats what youre asking
The piezo is a stacked actuator, the ones I use are 1.8uF they may get 2-3x larger. 200 ohm series resistor for the filter sometimes larger.
I'm not an expert but I was planning on using an opamp input, a gain stage(common emiter) maybe something else and then into a bias and output scheme and have the output feedback to the opamp and let it do the work. I could post up a circuit in a day or so I have been doing some simulations with bjts outputs but decided to step back when I realized there weren't any that could really handle what I need.
Heres a packaged amp I use PA15FL the price doubled when they were bought out by cirrus logic. The standard deviation noise brought back in is about 2ppm filtered at 200 Hz so this needs to be close (its in a closed loop system).
I will be sending out some pcbs next weeks so if I get something decent put together in the simulation I will panel it and not have to prototype my first trial
Again all thoughts are welcome I have no real experience designing these things just been reading up for the last couple months.
Popular high voltage devices listed
big insane TO-3 case MJ21194 & MJ21193
TO-220 case MJE15033 & MJE15032
TO-92 case BR423 & BF422
All of these are commonly available and all are 250 volt NPN & PNP devices that have a decent SOA. EVERYBODY uses these devices in literally hundreds of brands- it's crazy! MCM has the little ones right now but the change to stupid ROHS is messing up supply lines bad. Mouser is 19000 on back order due in AUGUST!!
big insane TO-3 case MJ21194 & MJ21193
TO-220 case MJE15033 & MJE15032
TO-92 case BR423 & BF422
All of these are commonly available and all are 250 volt NPN & PNP devices that have a decent SOA. EVERYBODY uses these devices in literally hundreds of brands- it's crazy! MCM has the little ones right now but the change to stupid ROHS is messing up supply lines bad. Mouser is 19000 on back order due in AUGUST!!
I would see if I could find a schematic/service manual for a EP2500 or 4000 - as an individual building just one you can't buy the components that go into cheap PA amps for $300
It may have AC coupling that requires some circuit mods but it would be way faster/cheaper to buy and mod a PA than designing and building from scratch (the outrageous power # are required to get the V swing even if the dissipation is only % of the PA's rating)
It may have AC coupling that requires some circuit mods but it would be way faster/cheaper to buy and mod a PA than designing and building from scratch (the outrageous power # are required to get the V swing even if the dissipation is only % of the PA's rating)
Last edited:
Hi,
the SOAR tells you what current & voltage the device can pass when Tc=25degC. There is a current limit, there is a power limit, there is a voltage limit and usually there is a second breakdown limit. This makes the graph look like 4lines one of which (second breakdown) is often curved.
There are usually extra short term SOAR superimposed over the DC SOAR.
All of these limits will require to be de-rated for actual operating Tc. ONsemi and others tell you how to do this.
Build it or buy it?
Like the other guy said- maybe buying is more useful- That PA08A I have list for like $186- would sell you one for half that including shipping. For $60 little more I can even provide the heat sink, rectifier and filter caps. Because it is an op-amp it may be run with large DC offsets and modulated- basically a modulatable power supply with a 150ma output.
A lot of off the shelf power amps are going to baulk at a capacitive load because so many (90%) are not very stable so have output inductors and are tested with resistive loads only. This opamp shows how well capacitive loads are driven in the data sheet sent.
Like the other guy said- maybe buying is more useful- That PA08A I have list for like $186- would sell you one for half that including shipping. For $60 little more I can even provide the heat sink, rectifier and filter caps. Because it is an op-amp it may be run with large DC offsets and modulated- basically a modulatable power supply with a 150ma output.
A lot of off the shelf power amps are going to baulk at a capacitive load because so many (90%) are not very stable so have output inductors and are tested with resistive loads only. This opamp shows how well capacitive loads are driven in the data sheet sent.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.