Ok, first I've done my searches but have not really come up with what I'm looking for.
Here is what I am trying to do with this amplifier.
Portable monitoring speaker for home built radios.
A portable, battery powered amplifier with low noise and low distortion. Output should be a clean 1-Watt into 8-Ohms preferably, regardless of the so called Wattage rating of the amplifier. Voltage is my friend and current is my enemy. Low quiescent current is a must but I do not mind using between one and four 6-volt Zinc Chloride (Heavy Duty) lantern batteries in series. The lantern batteries are from Dollar General and cost $2.00 each. Here are their specs. 7400 mAh @ 110-Ohm load with a 3.6-volt endpoint. I'm guessing that is about an average of a 40mAh current draw, so lets just say 185-hours of use at that current draw. I think that is realistic for an efficient speaker used for communications monitoring at say 70dB listening levels, with the ability to go louder cleanly when needed.
Total voltage gain for the amplifier is 1000. That includes of course the preamp. I plan on using between one and two OPA2132's that I already have on hand. The first stage will be a unity gain buffer with either one or three voltage gain stages after it. Small value resistors will be used in first stages for low thermal noise.
One of the problems I face is this. I really want to use a split supply because I do not see an elegant way to bias an op-amp for a unity-gain buffer with a single supply. I am not opposed to using a virtual ground at all but wouldn't I want to use a power amp that also uses a virtual ground? I do not really understand how to connect a preamp using a virtual ground to a power amp using a zero-volt ground. It even scares me connecting a radio that has it's signal ground earthed connected to the preamp with a virtual ground floating at say 6-volts.
I am also leaning heavily to class AB operation. Class D basically has an oscillator in it and it doesn't seem smart to me to have it in close vicinity of a very sensitive radio receiver.
I am pulling my hair out literally over this. I am totally against the LM386 "Noise Generator", just throwing that out there. I have looked seriously at the LM675 Power op-amp from National that looks like a stripped down LM1875 with lower quiescent current. I am also looking at the MAX9711 but the small SMT part brings up soldering concerns. Any help is greatly appreciated!!!
If I wasn't clear on anything I will gladly answer any questions. I can post a block diagram or schematic if needed.
Thank you all
Here is what I am trying to do with this amplifier.
Portable monitoring speaker for home built radios.
A portable, battery powered amplifier with low noise and low distortion. Output should be a clean 1-Watt into 8-Ohms preferably, regardless of the so called Wattage rating of the amplifier. Voltage is my friend and current is my enemy. Low quiescent current is a must but I do not mind using between one and four 6-volt Zinc Chloride (Heavy Duty) lantern batteries in series. The lantern batteries are from Dollar General and cost $2.00 each. Here are their specs. 7400 mAh @ 110-Ohm load with a 3.6-volt endpoint. I'm guessing that is about an average of a 40mAh current draw, so lets just say 185-hours of use at that current draw. I think that is realistic for an efficient speaker used for communications monitoring at say 70dB listening levels, with the ability to go louder cleanly when needed.
Total voltage gain for the amplifier is 1000. That includes of course the preamp. I plan on using between one and two OPA2132's that I already have on hand. The first stage will be a unity gain buffer with either one or three voltage gain stages after it. Small value resistors will be used in first stages for low thermal noise.
One of the problems I face is this. I really want to use a split supply because I do not see an elegant way to bias an op-amp for a unity-gain buffer with a single supply. I am not opposed to using a virtual ground at all but wouldn't I want to use a power amp that also uses a virtual ground? I do not really understand how to connect a preamp using a virtual ground to a power amp using a zero-volt ground. It even scares me connecting a radio that has it's signal ground earthed connected to the preamp with a virtual ground floating at say 6-volts.
I am also leaning heavily to class AB operation. Class D basically has an oscillator in it and it doesn't seem smart to me to have it in close vicinity of a very sensitive radio receiver.
I am pulling my hair out literally over this. I am totally against the LM386 "Noise Generator", just throwing that out there. I have looked seriously at the LM675 Power op-amp from National that looks like a stripped down LM1875 with lower quiescent current. I am also looking at the MAX9711 but the small SMT part brings up soldering concerns. Any help is greatly appreciated!!!
If I wasn't clear on anything I will gladly answer any questions. I can post a block diagram or schematic if needed.
Thank you all
The quiescent current of your amp will probably already be higher than 40 mA.
Are you sure you need that much gain for a radio? If you can get away with less gain, you will have to worry less about noise.
Coupling devices with single supply to others forces you to use AC coupling, i. e. capacitors in series with inputs and outputs. Once those are in place, there is no need to worry about connecting devices with different supply topologies to each other.
Hi
What kind of home built radios? I would tend to use the same Vcc of your radio design to drive your audio pwr amp, if you need more power for low ohms speakers BTL designs can be used. Using 12-14V nominal for Raw Vcc is nice for lots of reasons.
Usually the noise and distortion of the radio system is determined by the IF filters and AGC. The better radios implement additional AF filters as well as squelch circuitry to control noise. Using high power Hi Fi dual supply pwr amps will rarely buy anything but higher Iq and less battery time and worse of all, extra batteries. Using single supply AF amps biased at 1/2 Vcc and input DC blocking caps are nothing to fear or too complex to understand.
LM4752 perhaps?
What kind of home built radios? I would tend to use the same Vcc of your radio design to drive your audio pwr amp, if you need more power for low ohms speakers BTL designs can be used. Using 12-14V nominal for Raw Vcc is nice for lots of reasons.
Usually the noise and distortion of the radio system is determined by the IF filters and AGC. The better radios implement additional AF filters as well as squelch circuitry to control noise. Using high power Hi Fi dual supply pwr amps will rarely buy anything but higher Iq and less battery time and worse of all, extra batteries. Using single supply AF amps biased at 1/2 Vcc and input DC blocking caps are nothing to fear or too complex to understand.
LM4752 perhaps?
pacificblue
The quiescent current should be lower actually. I should have mentioned I currently have an amplifier that is single supply running on 12VDC. AD820AN preamp with a gain of 20 and a LM380 power amp with fixed gain of 50. If i remember right quiescent current was measured at around 30mA and I'm looking for more efficiency but noise is of more concern. The LM675 has a quiescent current of only 18mA.
Yes, I need that much gain. Actually that is low gain. I am coming straight off the detector but regenerative detectors have quite a bit of gain already. Probably somewhere in the neighborhood of 10-100k depending on design. Direct conversion receiver's audio section need much more gain because the detector does not supply it.
Ok, gotcha on the AC coupling. My main concern was referencing two different ground levels from the same power source. IE shorting the virtual ground out to real ground.
infinia
Regenerative receivers using FETs or BJTs. No IF filters or AGC in these designs. Very, very sensitive and very good audio quality. My main thing is DXing, the copying of extremely weak signals. For example receiving broadcast band AM transmissions from Saudi Arabia. I haven't done that but others have. So I do want very low noise so as not to add anything above the noise inherent in the detector.
If you know of a good way to bias an op-amp to half supply for a unity gain buffer please let me know. The way I have seen it done lowers the input impedance to the value of the resistor supplying 1/2 VCC to the positive input. I guess if it were say 1meg or above it wouldnt matter a whole lot.
The LM4752 looks interesting but for mono I would need a free op-amp channel to feed one of the inputs with an inverting signal.
Thank you again and all input is appreciated.
The quiescent current should be lower actually. I should have mentioned I currently have an amplifier that is single supply running on 12VDC. AD820AN preamp with a gain of 20 and a LM380 power amp with fixed gain of 50. If i remember right quiescent current was measured at around 30mA and I'm looking for more efficiency but noise is of more concern. The LM675 has a quiescent current of only 18mA.
Yes, I need that much gain. Actually that is low gain. I am coming straight off the detector but regenerative detectors have quite a bit of gain already. Probably somewhere in the neighborhood of 10-100k depending on design. Direct conversion receiver's audio section need much more gain because the detector does not supply it.
Ok, gotcha on the AC coupling. My main concern was referencing two different ground levels from the same power source. IE shorting the virtual ground out to real ground.
infinia
Regenerative receivers using FETs or BJTs. No IF filters or AGC in these designs. Very, very sensitive and very good audio quality. My main thing is DXing, the copying of extremely weak signals. For example receiving broadcast band AM transmissions from Saudi Arabia. I haven't done that but others have. So I do want very low noise so as not to add anything above the noise inherent in the detector.
If you know of a good way to bias an op-amp to half supply for a unity gain buffer please let me know. The way I have seen it done lowers the input impedance to the value of the resistor supplying 1/2 VCC to the positive input. I guess if it were say 1meg or above it wouldnt matter a whole lot.
The LM4752 looks interesting but for mono I would need a free op-amp channel to feed one of the inputs with an inverting signal.
Thank you again and all input is appreciated.
What about selectivity, local stations will blow out the detector. Using litz wire and air caps the loaded Q will be? remember antenna's are loads on preselect filters/tuners too.
Why a unity gain stage, ? since all your gain is post detection, I would look for the lowest voltage noise preamp with gain following the detector.
This interesting to me , I could help you quicker if you care to point me to a typical receiver you are referencing as well as what you have built already.
Use only 1/2 of the stereo pair then, For 1W with an 8 ohm speaker and your batteries running down to say 9V or so, you'll need to boost up to BTL. The LM4752 is what you want! It's a modern low noise power amp with Iq less than 8 mA in a great package. Look at it closer or generate a list of complete requirements.
Why a unity gain stage, ? since all your gain is post detection, I would look for the lowest voltage noise preamp with gain following the detector.
This interesting to me , I could help you quicker if you care to point me to a typical receiver you are referencing as well as what you have built already.
Most op-amp noise is a function of input impedance, so your bias network actually needs to form a lowish Thevinen source. There is an optimum input impedance to balance the current noise to the voltage noise.
Use only 1/2 of the stereo pair then, For 1W with an 8 ohm speaker and your batteries running down to say 9V or so, you'll need to boost up to BTL. The LM4752 is what you want! It's a modern low noise power amp with Iq less than 8 mA in a great package. Look at it closer or generate a list of complete requirements.
No. Superhets have replaced regens and super-regens for a reason. They are superior in most respects other than in the number of components they employ.
w
Well I don't want to get into a whole radio discussion but I will give brief info. A regenerative has very high Q because of the positive feedback (Q multiplication) enough to cut sidebands. As for locals, either use a tunable loop antenna, wave trap, or double tuned circuit if they are a problem.
The amp is to be separate from the radio as to be able to be used between various equipment.
Unity gain buffer to provide a very high impedance to not load down the detector and hurt selectivity.
I will look heavily at the amp you recommended.
Thank you for the input.
The amp is to be separate from the radio as to be able to be used between various equipment.
Unity gain buffer to provide a very high impedance to not load down the detector and hurt selectivity.
I will look heavily at the amp you recommended.
Thank you for the input.
Yes superhets have replaced them for various reasons. Mainly ease of consumer use and simplifying selectivity. Superhets do have much more distortion on AM but not FM. The direct conversion receiver beats a superhet when properly designed. They were actually designed to replace superhets a long time ago but due to the limited technology of the day it was impractical. Your cell phone has a direct conversion receiver inside of it as do SDR software defined radios that are going to replace everything.
Well there is not really any comparison to what modern cellphones use and what you are doing ie digital baseband filters are equal to an IF filter.
Can you post yer schematics?
Can you post yer schematics?
Well I do not have a schematic on hand but here is one of the radios I studied. It is an older design but basically it explains the principles involved.
http://www.arrl.org/files/file/Technology/tis/info/pdf/9811qex026.pdf
http://www.arrl.org/files/file/Technology/tis/info/pdf/9811qex026.pdf
OK I remember seeing that Rx a long long time ago, looks like fun. Hours of tweaking everything LOL. Did you say you have built this retuned for MW? I have a great coil for the AM broadcast band, A beautiful air coil, factory made with litz wire and lots of taps.
Look at the LT-1028 app note for pre-amp noise etc. should be a big improvement.
Look at the LT-1028 app note for pre-amp noise etc. should be a big improvement.
Last edited:
see attached fixed AF buffer
bias is set at 4.5 V with about 11K input replacing 10K pot. The input bias supply is isolated and filtered by the 2K RC network. The output gain is lowered here because the pwr amp makes up the additional gain. Volume control is moved to input of the pwr amp. You might need to change the fixed gain once you get up and running and see where the volume control likes to be. .
Probably the next big improvement could be made by using a steeper audio CLC filter instead of the R5*C.
bias is set at 4.5 V with about 11K input replacing 10K pot. The input bias supply is isolated and filtered by the 2K RC network. The output gain is lowered here because the pwr amp makes up the additional gain. Volume control is moved to input of the pwr amp. You might need to change the fixed gain once you get up and running and see where the volume control likes to be. .
Probably the next big improvement could be made by using a steeper audio CLC filter instead of the R5*C.
Attachments
Last edited:
here is a good read on audio filters to replace between your front end and the fixed AF amp. His design is not a simple one like yours ( note the all important LO Local Oscillator is glossed over) http://www.arrl.org/files/file/Technology/tis/info/pdf/9208019.pdf
edit> I forgot the series DC block capacitor in series with R10 on the attached marked up AF buffer above.
edit> I forgot the series DC block capacitor in series with R10 on the attached marked up AF buffer above.
Last edited:
A regen is not a DC receiver.
It is a mistake to imagine that fidelity in broadcast AM is of importance. The channel spacing is only 10kHz at best, and with 2 sidebands and a guard band the baseband audio can only occupy ~4.5kHz, 5kHz at absolute maximum. The performance of a regen will not be noticeably better than that of a well designed superhet anyway. You have to remember that the author is a regen enthusiast, and likely to exaggerate it's qualities.
Even with careful attention to design a regen will not exceed the performance of a superhet and will fall short in many circumstances. They're cool for a bit of fun, I'm not saying don't build one, what I am saying is that it will not exceed the performance of a superhet and in modern band conditions in many cases it will fall short. You can install traps in your antenna, but how many, and how portable will it be then? Yes, you can add dual- or triple-TRF stages, but they require ganged caps. This adds complexity, makes setup more difficult and increases the likelihood of unwanted feedback, making layout more problematic. This is just one of the reasons why superhets are attractive, because the gain is distributed at more than one frequency. A basic regen is, however, easier to build than a superhet or even a DC receiver, although much of the effort of building a DC receiver has been reduced with the advent of DDS. You can get a PCB for a crystal controlled DDS.
You don't really need a low noise audio amp. The NF of the radio, which is what determines the MDS, is dominated by the front end, with maybe 86dB of gain.
Here's a low-noise amp from Hayward & DeMaw, Solid State Design for the Radio Amateur, it's intended for a DC receiver, it's a bit overkill, it has ~100dB gain. Obviously it's intended to drive 2k, but it's trivial to add a Darlington or Mosfet buffer. A typical diode product detector has an output impedance of ~50R, so there would normally be a txfrmr at the input.
Since the output of a regen is frequently on the order of some 100s of mV and the circuit shown already has some audio amplification, you will probably be better with one of these:-
The Motorola devices are obsolete, but still obtainable in the US.
I thoroughly recommend the book, ISBN 0-87259-040-2.
w
It is a mistake to imagine that fidelity in broadcast AM is of importance. The channel spacing is only 10kHz at best, and with 2 sidebands and a guard band the baseband audio can only occupy ~4.5kHz, 5kHz at absolute maximum. The performance of a regen will not be noticeably better than that of a well designed superhet anyway. You have to remember that the author is a regen enthusiast, and likely to exaggerate it's qualities.
Even with careful attention to design a regen will not exceed the performance of a superhet and will fall short in many circumstances. They're cool for a bit of fun, I'm not saying don't build one, what I am saying is that it will not exceed the performance of a superhet and in modern band conditions in many cases it will fall short. You can install traps in your antenna, but how many, and how portable will it be then? Yes, you can add dual- or triple-TRF stages, but they require ganged caps. This adds complexity, makes setup more difficult and increases the likelihood of unwanted feedback, making layout more problematic. This is just one of the reasons why superhets are attractive, because the gain is distributed at more than one frequency. A basic regen is, however, easier to build than a superhet or even a DC receiver, although much of the effort of building a DC receiver has been reduced with the advent of DDS. You can get a PCB for a crystal controlled DDS.
You don't really need a low noise audio amp. The NF of the radio, which is what determines the MDS, is dominated by the front end, with maybe 86dB of gain.
Here's a low-noise amp from Hayward & DeMaw, Solid State Design for the Radio Amateur, it's intended for a DC receiver, it's a bit overkill, it has ~100dB gain. Obviously it's intended to drive 2k, but it's trivial to add a Darlington or Mosfet buffer. A typical diode product detector has an output impedance of ~50R, so there would normally be a txfrmr at the input.
Since the output of a regen is frequently on the order of some 100s of mV and the circuit shown already has some audio amplification, you will probably be better with one of these:-
The Motorola devices are obsolete, but still obtainable in the US.
I thoroughly recommend the book, ISBN 0-87259-040-2.
w
waki those 60's amp designs are pretty pathetic. not surprising most of those parts are obsolete. It would be a mistake to build them esp the 1st. one. 100 db gain co-located is a waiting oscillator nightmare. The OP is wisely updating an ancient design using 2 stages of opamp gain on seperate PCBs. I recommend him a modern low Iq monolythic audio power amp and you offer him this. LOL
Direct conversion Rx's have a few advantages over superhets esp involving problem AGCs and CW DXing, no narrow band IF group delay distortion., etc. Signal Fidelity is a well known trait among the amatuer radio folks.
Direct conversion Rx's have a few advantages over superhets esp involving problem AGCs and CW DXing, no narrow band IF group delay distortion., etc. Signal Fidelity is a well known trait among the amatuer radio folks.
wakibaki
OK, fine. Regens suck, end of story. This is to be a separate audio amplifier that can be used with various equipment.
infinia
You definitely have me hooked on the single supply. After looking very heavily at the LM4752 I also ran across the LM4950. Higher quiescent current but built for 12VDC, externally adjusted gain, and simple BTL for single 8-Ohm speaker. The numbers look good to me.
http://www.national.com/ds/LM/LM4950.pdf
Look at page 6, figure labeled 200470A9 for THD+N for BTL, unitygain, 8-Ohm load @12VDC. Tell me what you think.
I am definitely homing in on the power amp portion of the project!!!
Thanks
OK, fine. Regens suck, end of story. This is to be a separate audio amplifier that can be used with various equipment.
infinia
You definitely have me hooked on the single supply. After looking very heavily at the LM4752 I also ran across the LM4950. Higher quiescent current but built for 12VDC, externally adjusted gain, and simple BTL for single 8-Ohm speaker. The numbers look good to me.
http://www.national.com/ds/LM/LM4950.pdf
Look at page 6, figure labeled 200470A9 for THD+N for BTL, unitygain, 8-Ohm load @12VDC. Tell me what you think.
I am definitely homing in on the power amp portion of the project!!!
Thanks
there are several million table radios out there that put out less that 250 mw. 1/4 watt. limit your audio bandwidth to 200-4khz. You need gain but you need audio compression as well. If you design a wide bandwidth amp you will be wasting power. I am putting the final touches on a hi fi audio amp. The output noise is an average of about 400 micro volts. Not that great by todays standards. I can bearly hear the noise if I put my ear next to the speakers. Thats about 2 micro watts of audiable output power. So several millivolts/milliwatts will be audible esp with headphones. Are you sure you need that much audio gain? A simple emitter follower following any single supply op amp should work. Look for micro power op amps that are made to run on single supply voltages ie LM124 as an example
Hi Squareguy
Yes I like single supplies for mobile gear too. These newer power amps are real nice in that they include all the input biasing internally.
The LM4950 looks good esp if you are going single channel at 12V. BTL is easier to use on this because of unity gain stabilty. Also the output RC snubbers are NOT required which is a bonus. I still like the higher power capability of the 52 version for a general purpose stereo amp tho. Either one is good.
Yes I like single supplies for mobile gear too. These newer power amps are real nice in that they include all the input biasing internally.
The LM4950 looks good esp if you are going single channel at 12V. BTL is easier to use on this because of unity gain stabilty. Also the output RC snubbers are NOT required which is a bonus. I still like the higher power capability of the 52 version for a general purpose stereo amp tho. Either one is good.
yes there are many reasons to have a narrow final BW ! and 200 Hz to 4000 Hz is about right. It's best to limit the bandwith via interstage filtering and wideband amps. I'm not aware of any power penalty for using wideband op amps tho. There are many newer op amps that are far superior for audio than the old GP workhorses tho. I think a worthy goal is to hear no noise with gain cranked up and no antenna connected.
Attachments
You are joking, aren't you infinia? You've got a filter with 4k bandwidth and it's being recommended for hi-fidelity? This just reinforces my earlier comments about AM radio.
The circuits I posted are entirely adequate to the task, they come with an impeccable pedigree, and they do not date from the 60's. Yes, 100dB is a lot of gain at baseband, but I didn't suggest using that one. I chose to post these circuits because the reputations W7ZOI and W1FB enjoy in the amateur community will permit any serious amateur to use them with confidence.
Radio has been around for a long time, as is evidenced by the existence of the regen receiver. Your suggestions for modern monolithic devices will not result in noticeably improved performance, they are the same kind of overoptimistic, unrealistic, fantastical approach that calls 200-4000Hz hi-fi, and leads ultimately to the purchase of silver wire mains cables.
w
The circuits I posted are entirely adequate to the task, they come with an impeccable pedigree, and they do not date from the 60's. Yes, 100dB is a lot of gain at baseband, but I didn't suggest using that one. I chose to post these circuits because the reputations W7ZOI and W1FB enjoy in the amateur community will permit any serious amateur to use them with confidence.
Radio has been around for a long time, as is evidenced by the existence of the regen receiver. Your suggestions for modern monolithic devices will not result in noticeably improved performance, they are the same kind of overoptimistic, unrealistic, fantastical approach that calls 200-4000Hz hi-fi, and leads ultimately to the purchase of silver wire mains cables.
w
- Status
- Not open for further replies.