wine&dine said:
I agree, I don't think a regulator is necessary. I just didn't think it would be good to run the motor on 12V when it was originally designed for 230 which is what I thought you were getting at.
I follow that rough schematic (amazing considering) it helps a lot and I've started to source some ICs already, thanks a bunch guys. Hopefully you can help me (and I will have) with more specific questions
Also, I started doing some very rough calculations. Not worrying about all the power supply required to run everything yet, but I figure a simple power op-amp tagged on after the schematic on the second page would work? I was looking into:
LT1210 and AD8392
outputting ~ 20V at ~300mA
then to approx. 9:1 transformer. Whole system would run approx. 6W (double the required). I think I'm simiplfying way too much, so tell me if I'm way off base or if I'm at least scratching the surface. I figured a quad op amp setup could run a buffer stage, then an active LPF and then two amp stages on the single quad amp
LT1210 and AD8392
outputting ~ 20V at ~300mA
then to approx. 9:1 transformer. Whole system would run approx. 6W (double the required). I think I'm simiplfying way too much, so tell me if I'm way off base or if I'm at least scratching the surface. I figured a quad op amp setup could run a buffer stage, then an active LPF and then two amp stages on the single quad amp
For personal reference and anyone researching/designing as well:
http://www.vinylparadise.com/ttable/hercules/hercules.htm
http://cgi.ebay.ca/HERCULES-POWER-S...oryZ3283QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
translated:
http://translate.google.com/transla...&hl=en&ie=UTF-8&oe=UTF-8&prev=/language_tools
http://translate.google.com/transla...9bf1bdfcc39a6&langpair=zh-CN|en&hl=en&ie=UTF8
http://www.vinylparadise.com/ttable/hercules/hercules.htm
http://cgi.ebay.ca/HERCULES-POWER-S...oryZ3283QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
translated:
http://translate.google.com/transla...&hl=en&ie=UTF-8&oe=UTF-8&prev=/language_tools
http://translate.google.com/transla...9bf1bdfcc39a6&langpair=zh-CN|en&hl=en&ie=UTF8
>I think I'm simiplfying way too much
Au contraire, I think you should simplify *further*. Look for a cheap chip amp designed for bridged loads at 12V (car radio output). For example, the TDA1519C sells for < EUR 5 around here, only needs one external decoupling cap when wired as a bridge amp, and can put out some 20W. I'd even try reducing the filter to a simple R/C at first, you can always add another stage with just an emitter follower.
Those links of yours make me wonder if I'm in the wrong business - 150 quid for couple of watts at line frequency ... yikes.
Au contraire, I think you should simplify *further*. Look for a cheap chip amp designed for bridged loads at 12V (car radio output). For example, the TDA1519C sells for < EUR 5 around here, only needs one external decoupling cap when wired as a bridge amp, and can put out some 20W. I'd even try reducing the filter to a simple R/C at first, you can always add another stage with just an emitter follower.
Those links of yours make me wonder if I'm in the wrong business - 150 quid for couple of watts at line frequency ... yikes.
wine&dine said:
Thanks, I will look into a simpler chip amp and a opamp active filter should be fine to work with, I could even do a 5th stage butterworth with a quad chip
...and here I was going to comment on the power supply on ebay, but I thought I was being ignorant thinking that it was WAY overpriced. I can almost Identify most of the setup, but can't find any really high def. pictures... only chip I can pick out after the crystal stage is an LM324 - a LP filter or pre-amp stage perhaps?, seems like that setup is similar to what others have suggested here.
Quick Edit - thanks wine&dine
Part - TDA1519CU
3.59 at digikey
datasheet with BTL setup
http://www.semiconductors.philips.com/acrobat_download/datasheets/TDA1519C_4.pdf
amazing...
No quad opamp needed. No t even one ...
Take a look at this ... the input is a 12V square from the 4040, V2 is the supply, V3 is the TDA1519 internal bias, R4 is the TDA's 2x60k inputs in parallel. Output trace (red) will be amplified x200 (40dB x2 by bridge) by the TDA1519. Make R5 a trimpot and Bob's your uncle.
Take a look at this ... the input is a 12V square from the 4040, V2 is the supply, V3 is the TDA1519 internal bias, R4 is the TDA's 2x60k inputs in parallel. Output trace (red) will be amplified x200 (40dB x2 by bridge) by the TDA1519. Make R5 a trimpot and Bob's your uncle.
Attachments
Did you mean make R3 a trim pot?
Also, if I understand the configuration correctly, the BTL setup acts as an amplifier with a differential setup (?)
Also, I would still need a transformer correct? would this cause any problems for the chip amp because of the type of the load that the transformer is?
sorry if I'm misunderstanding anything
Also, if I understand the configuration correctly, the BTL setup acts as an amplifier with a differential setup (?)
Also, I would still need a transformer correct? would this cause any problems for the chip amp because of the type of the load that the transformer is?
sorry if I'm misunderstanding anything
>Did you mean make R3 a trim pot?
Sorry, I mean make R2 a trimpot, with R3 connected to the wiper, to set the output voltage.
>Also, if I understand the configuration correctly,
>the BTL setup acts as an amplifier with a differential setup (?)
I don't understand the question. You feed it (bias+50mV), one output goes to (6+5)V, the other to (6-5)V, or 10V across the load.
100 mVss -> 20Vss.
>Also, I would still need a transformer correct?
Yep. Use 230/6V. 6Veff=17Vss, which the TDA1519 can do at 12V. For 17Vss it needs 85mVss in.
>would this cause any problems for the chip amp
>because of the type of the load that the transformer is?
Nope.
Sorry, I mean make R2 a trimpot, with R3 connected to the wiper, to set the output voltage.
>Also, if I understand the configuration correctly,
>the BTL setup acts as an amplifier with a differential setup (?)
I don't understand the question. You feed it (bias+50mV), one output goes to (6+5)V, the other to (6-5)V, or 10V across the load.
100 mVss -> 20Vss.
>Also, I would still need a transformer correct?
Yep. Use 230/6V. 6Veff=17Vss, which the TDA1519 can do at 12V. For 17Vss it needs 85mVss in.
>would this cause any problems for the chip amp
>because of the type of the load that the transformer is?
Nope.
wine&dine said:
lol, thats what I meant I duplicated your model and switched the resistor numbers on mine, I understand the logic.
wine&dine said:
I understand how it creates the reference at half of the Vp (power supply input (I think)... but then how could it create 17Vss (we say Vpp, confused me for a second) from 12V?? wouldn't it need to reference 8.5V as zero, then you need at least a 17V at Vp? or more conservatively (without saturation) higher then 17V at Vp. Because to get the 17Vss it would be one at (8.5-8.5)V and the other at (8.5+8.5)V.... with 12V at Vp, it would go (6-6)V and (6+6)V max.... or am I missing something again?
I understand the bias + 85mV ---> 17Vss just fine.
> 17Vss (we say Vpp, confused me for a second)
sorry, brain not re-wired 100% to English before 10am ;-)
> how could it create 17Vpp from 12V?
It's a bridge-tied load - BOTH ends of the load are driven.
Positive input: Output 1 goes to (6V + 8.5/2 V) = 10.25V (< 12V - Vsat, OK?); Output 2 goes to (6V - 8.5/2 V) = 1.75V (> Vsat, OK?). In this half of the cycle, you therefore get 8.5V peak (Out1-Out2).
Negative input reverses the roles of O1 and O2, so you get (-8.5V) peak on the load. Between the positive and negative peaks, that's 17Vpp.
Clear now?
sorry, brain not re-wired 100% to English before 10am ;-)
> how could it create 17Vpp from 12V?
It's a bridge-tied load - BOTH ends of the load are driven.
Positive input: Output 1 goes to (6V + 8.5/2 V) = 10.25V (< 12V - Vsat, OK?); Output 2 goes to (6V - 8.5/2 V) = 1.75V (> Vsat, OK?). In this half of the cycle, you therefore get 8.5V peak (Out1-Out2).
Negative input reverses the roles of O1 and O2, so you get (-8.5V) peak on the load. Between the positive and negative peaks, that's 17Vpp.
Clear now?
I finished a complete schematic, I wont bother routing it I think I'm going to use a project/prototype board... any and all criticisms are welcome... I'm still picking components in stock at digikey and have to see if I have any transformers lying around that I can use. Thanks for everyones help.
Voltage too high on LM7812
I think the connection between the transformer and the TDA1519 outputs is incorrect. From the data sheet the outputs are designed to drive a load into ground, and not into each other. What you would probably need to do is drive them into the ends of a center tapped primary with the center tap attached to ground.
The two inputs (pins 9 and 1) also need to complement each other so that the when input 1 is high, input 2 is low. As it stands your driving both ends with the same voltage.
FWIW: I don't think this will work. You might be better off consider using a a pair of mosfets (IRFZ24N) attached to the primary using a push-pull configuration and use a 555 RC oscillator set up with produce a sine wave output and a couple of bipolar transistors totem pole pairs ( 2N3904 & 2N3906) to drive the mosfet gates. Using a crystal and a couple of logic ICs seems a bit overkill to generate a 50 hz signal.
A pair of low voltage (55V) high current mosfets are inexpensive and would be able to handle a bigger load and take more punishment then the stereo amp . IRFZ24N (TO-220, 17A Pd=45W) cost about 52 cents each from digikey.com
Keep in mind that because you will be driving the transformer with a slow sine wave, there will be a lot of power wasted in the mosfets (or the stereo amp). You will probably need a much bigger input transformer to produce the 3 to 5 VAC at 220 AC @ 50 Hz. Also don't drive the transformer through the regulator. Take the current from the rectifier output instead. If you make a mistake, you could over load and damage the regulator. If you use a 555 RC oscillator instead, you can probably forego the regulator and just use a small coupling cap connected close to the 555 power pins, as I recall, the 555 can operate at 18V. At such a low frequency, input power noise isn't going to be a problem. (Astable oscillator calc: http://www.csgnetwork.com/ne555timer2calc.html)
Alternatively for a more professional design, look at this application note showing a DC/AC inverter using the HIP4082.
http://powerdesigners.com/InfoWeb/forums/conv/uploads/AN9611.pdf
Good Luck
I think the connection between the transformer and the TDA1519 outputs is incorrect. From the data sheet the outputs are designed to drive a load into ground, and not into each other. What you would probably need to do is drive them into the ends of a center tapped primary with the center tap attached to ground.
The two inputs (pins 9 and 1) also need to complement each other so that the when input 1 is high, input 2 is low. As it stands your driving both ends with the same voltage.
FWIW: I don't think this will work. You might be better off consider using a a pair of mosfets (IRFZ24N) attached to the primary using a push-pull configuration and use a 555 RC oscillator set up with produce a sine wave output and a couple of bipolar transistors totem pole pairs ( 2N3904 & 2N3906) to drive the mosfet gates. Using a crystal and a couple of logic ICs seems a bit overkill to generate a 50 hz signal.
A pair of low voltage (55V) high current mosfets are inexpensive and would be able to handle a bigger load and take more punishment then the stereo amp . IRFZ24N (TO-220, 17A Pd=45W) cost about 52 cents each from digikey.com
Keep in mind that because you will be driving the transformer with a slow sine wave, there will be a lot of power wasted in the mosfets (or the stereo amp). You will probably need a much bigger input transformer to produce the 3 to 5 VAC at 220 AC @ 50 Hz. Also don't drive the transformer through the regulator. Take the current from the rectifier output instead. If you make a mistake, you could over load and damage the regulator. If you use a 555 RC oscillator instead, you can probably forego the regulator and just use a small coupling cap connected close to the 555 power pins, as I recall, the 555 can operate at 18V. At such a low frequency, input power noise isn't going to be a problem. (Astable oscillator calc: http://www.csgnetwork.com/ne555timer2calc.html)
Alternatively for a more professional design, look at this application note showing a DC/AC inverter using the HIP4082.
http://powerdesigners.com/InfoWeb/forums/conv/uploads/AN9611.pdf
Good Luck
Re: Voltage too high on LM7812
First of all, let me thank you for your kind response and insight into my design. Its too bad I already ordered everything from digikey and will start construction tonight or tomorrow, I do believe that it will work though.
the transformer I am using can be center tapped to ground if need be, but in BTL configuration, the data sheet specifies there is no need, so I think this should be ok.
Not sure what you mean, but according to the data sheet and what I went through with wine&dine the amp bias to half the input voltage so it can operate from a single supply, if you look at the diagram above it shows the same.
The IC+crystal method has been used numerous times - including a few commercial designs. I've searched and was suggested in this thread that RC oscillators are not accurate enough, and the crystal configuration is really easy to implement so why not go for it. Its too late for me to change my drive stage again because I already have parts in the mail, but if I blow the amp I will consider a mosfet drive stage - I'm being lazy, chip amps are quick and easy. Also I looked through all the specs onthe amp and other components before ordering, if I've done it correctly I won't need more power and nothing will blow/burn out.
I really appreciate the extra links. I think the more sophisticated design is overkill. I will drive the transformer amp straight from the rectifier instead, I think this is a good idea, thanks!
Will let anyone following along know how it turns out in a few days (hopefully)
First of all, let me thank you for your kind response and insight into my design. Its too bad I already ordered everything from digikey and will start construction tonight or tomorrow, I do believe that it will work though.
TechGuy said:
the transformer I am using can be center tapped to ground if need be, but in BTL configuration, the data sheet specifies there is no need, so I think this should be ok.
An externally hosted image should be here but it was not working when we last tested it.
TechGuy said:
Not sure what you mean, but according to the data sheet and what I went through with wine&dine the amp bias to half the input voltage so it can operate from a single supply, if you look at the diagram above it shows the same.
TechGuy said:
The IC+crystal method has been used numerous times - including a few commercial designs. I've searched and was suggested in this thread that RC oscillators are not accurate enough, and the crystal configuration is really easy to implement so why not go for it. Its too late for me to change my drive stage again because I already have parts in the mail, but if I blow the amp I will consider a mosfet drive stage - I'm being lazy, chip amps are quick and easy. Also I looked through all the specs onthe amp and other components before ordering, if I've done it correctly I won't need more power and nothing will blow/burn out.
TechGuy said:
I really appreciate the extra links. I think the more sophisticated design is overkill. I will drive the transformer amp straight from the rectifier instead, I think this is a good idea, thanks!
Will let anyone following along know how it turns out in a few days (hopefully)
>Not sure what you mean, but according to the data sheet and what I went through with wine&dine the amp bias to half the input voltage so it can operate from a single supply, if you look at the diagram above it shows the same.
Your right, I didn't notice that the one input was inverting while the other is non-inverting. I thought they were both the same.
>I've searched and was suggested in this thread that RC oscillators are not accurate enough, and the crystal configuration is really easy to implement so why not go for it.
Commerical AC power frequency varies (at least what I have monitored from time to time by about 1%). I doubt your device needs to be dead on accurate.
Your right, I didn't notice that the one input was inverting while the other is non-inverting. I thought they were both the same.
>I've searched and was suggested in this thread that RC oscillators are not accurate enough, and the crystal configuration is really easy to implement so why not go for it.
Commerical AC power frequency varies (at least what I have monitored from time to time by about 1%). I doubt your device needs to be dead on accurate.
Re: Re: Voltage too high on LM7812
>Using a crystal and a couple of logic ICs seems a
>bit overkill to generate a 50 hz signal.
He wants a stable 50 Hz with no fuss - 2 cheap 4xxx chips, so what? I wonder why nobody has suggested an 8-pin PIC controller for the job yet, it also costs pennies & has fewer pins to get wrong - but the poster would be asked to program one
>would be able to handle a bigger load and
>take more punishment then the stereo amp
3 Watt synchronous AC motors don't punish.
>slow sine wave, there will be a lot of power wasted
Frequency is irrelevant. At 12VDC to amp, 6Vrms to trafo, efficiency is ~50%: feed 5W to the trafo (3W + losses), dissipate 5W. A lot?
> don't drive the transformer through the regulator.
> you could over load and damage the regulator
Au contraire, use a current limiting regulator to protect the *amp*. An LM317 supplies up to 1.5A = 18W DC at 12V, allowing our little amp to put out 9W AC at 6Vrms. Plus, the '317 has thermal in addition to current limiting.
>the transformer I am using can be center tapped to ground if need be,
>but in BTL configuration, the data sheet specifies there is no need
Not only is there no "need" - the CT terminal is always at 1/2 V+ bias, connecting it to ground will create a short!
BTL means the load MUST be FLOATING.
>Using a crystal and a couple of logic ICs seems a
>bit overkill to generate a 50 hz signal.
He wants a stable 50 Hz with no fuss - 2 cheap 4xxx chips, so what? I wonder why nobody has suggested an 8-pin PIC controller for the job yet, it also costs pennies & has fewer pins to get wrong - but the poster would be asked to program one
>would be able to handle a bigger load and
>take more punishment then the stereo amp
3 Watt synchronous AC motors don't punish.
>slow sine wave, there will be a lot of power wasted
Frequency is irrelevant. At 12VDC to amp, 6Vrms to trafo, efficiency is ~50%: feed 5W to the trafo (3W + losses), dissipate 5W. A lot?
> don't drive the transformer through the regulator.
> you could over load and damage the regulator
Au contraire, use a current limiting regulator to protect the *amp*. An LM317 supplies up to 1.5A = 18W DC at 12V, allowing our little amp to put out 9W AC at 6Vrms. Plus, the '317 has thermal in addition to current limiting.
>the transformer I am using can be center tapped to ground if need be,
>but in BTL configuration, the data sheet specifies there is no need
Not only is there no "need" - the CT terminal is always at 1/2 V+ bias, connecting it to ground will create a short!
BTL means the load MUST be FLOATING.
Don't worry wine&dine, I actually haven't strayed from the design we came up with. I am half way there, I have 50.0000Hz saw tooths at the moment. Amazing precision out of a simple circuit. I am waiting for digikey to ship me some parts their automatic bin picker screwed up. I am very grateful, it is coming along very nicely!
>3 Watt synchronous AC motors don't punish.
I doubt that device's peak power input is only 3 watts. Motors have much higher power input requirements during startup. If the device has a step down transformer, its possible that the transformer loss might exceed a couple of watts alone. There isn't much wiggle room if the device need more than just 3 watts.
>Frequency is irrelevant. At 12VDC to amp, 6Vrms to trafo, efficiency is ~50%: feed 5W to the trafo (3W + losses), dissipate 5W. A lot?
Transistor conduction losses are high when they operate below saturation. For much of the cycle of the sinewave, the output transistors of the amp will be operating below saturation. thus causing significant power loss.
> Au contraire, use a current limiting regulator to protect the *amp*. An LM317 supplies up to 1.5A = 18W DC at 12V, allowing our little amp to put out 9W AC at 6Vrms. Plus, the '317 has thermal in addition to current limiting.
Whatever, dude. Maybe I am reading this wrong, but you seem to me as if this is some sort of deathmatch challege. I see your point but you don't need to be condescending about it.
I doubt that device's peak power input is only 3 watts. Motors have much higher power input requirements during startup. If the device has a step down transformer, its possible that the transformer loss might exceed a couple of watts alone. There isn't much wiggle room if the device need more than just 3 watts.
>Frequency is irrelevant. At 12VDC to amp, 6Vrms to trafo, efficiency is ~50%: feed 5W to the trafo (3W + losses), dissipate 5W. A lot?
Transistor conduction losses are high when they operate below saturation. For much of the cycle of the sinewave, the output transistors of the amp will be operating below saturation. thus causing significant power loss.
> Au contraire, use a current limiting regulator to protect the *amp*. An LM317 supplies up to 1.5A = 18W DC at 12V, allowing our little amp to put out 9W AC at 6Vrms. Plus, the '317 has thermal in addition to current limiting.
Whatever, dude. Maybe I am reading this wrong, but you seem to me as if this is some sort of deathmatch challege. I see your point but you don't need to be condescending about it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.