I recently purchased a nice little amplifier to use at my PC.
Came across an Eltax (or Tanget) AMP-50 which has a very nice looking interior for an amp of this price, especially what I paid second hand for it. Bought it thinking that if it did not work out, just the toroidal transformer together with the connectors and housing is worth it and I can diy something inside 😀
Image from Google, but you can see it is built nice and modular.
Well the amp runs great, and I like it a lot! But it has a constant hiss on both channels that I would like to see reduced.
Hopefully you can help me figure out in what part it is located first of all. What I noticed:
This leads me to believe that the noise is being generated in the poweramp side of the circuit, as volume control is done in the preamp and going into 'standby' lowers it.
LUCKILY the schematics are available 😀
My first 'instinct' was to lower the gain somehow. I do not run at high volumes and usually when putting it on, I turn down the volume from the initial position anyway. Also during my use it does not run hot the slightest.
Any help is appreciated 🙂 I would love to learn more
Came across an Eltax (or Tanget) AMP-50 which has a very nice looking interior for an amp of this price, especially what I paid second hand for it. Bought it thinking that if it did not work out, just the toroidal transformer together with the connectors and housing is worth it and I can diy something inside 😀
Image from Google, but you can see it is built nice and modular.
Well the amp runs great, and I like it a lot! But it has a constant hiss on both channels that I would like to see reduced.
Hopefully you can help me figure out in what part it is located first of all. What I noticed:
- Disconnecting all sources does not change it
- Increasing or lowering the volume does not change the volume of the hiss (white noise like)
- When no signal is fed, some kind of 'mute' kicks in, and the hiss lowers a bit
This leads me to believe that the noise is being generated in the poweramp side of the circuit, as volume control is done in the preamp and going into 'standby' lowers it.
LUCKILY the schematics are available 😀
My first 'instinct' was to lower the gain somehow. I do not run at high volumes and usually when putting it on, I turn down the volume from the initial position anyway. Also during my use it does not run hot the slightest.
Any help is appreciated 🙂 I would love to learn more
Last edited:
If you remove the preamp signal, Rout and Lout, so the amp is on its own, what happens to the noise?
New members are moderated, and it appears that you edited your opening post and it got stuck. Fixed now!
As you say, a nice little amp 🙂
Something for you to start with... I suspect the noise will be coming from the front end rather than the power amplifier but you should be able to prove that by isolating the two, but you have to do that correctly to avoid problems .
Having said that there appears to be an error in the diagram. Q3 and Q5 on the power amp input have their bases' floating, there is no path to ground to bias them. I was going to suggest lifting the left hand side of C1 and then connecting that lifted end of the cap to ground so that the input to the amp is 'shorted' by the cap.
I think you will find there is a resistor fitted from R1 to ground that is omitted from the diagram.
And ultimately if I'm honest, well I suspect you won't reduce the hiss as it will be inherent to the design. But its nice to prove where its coming from.
Something for you to start with... I suspect the noise will be coming from the front end rather than the power amplifier but you should be able to prove that by isolating the two, but you have to do that correctly to avoid problems .
Having said that there appears to be an error in the diagram. Q3 and Q5 on the power amp input have their bases' floating, there is no path to ground to bias them. I was going to suggest lifting the left hand side of C1 and then connecting that lifted end of the cap to ground so that the input to the amp is 'shorted' by the cap.
I think you will find there is a resistor fitted from R1 to ground that is omitted from the diagram.
And ultimately if I'm honest, well I suspect you won't reduce the hiss as it will be inherent to the design. But its nice to prove where its coming from.
That could be destructive if the power amp input doesn't see a low impedance path to ground. It also looks like the diagram is in error.
Q3 and Q5 are AC coupled so I don't expect issues but connecting the inputs; C1 positive lead, to ground will be belt and braces. I expect the noise is from IC3 or a noisy 4558.
Yes, AC coupled. What is missing on the diagram is a DC path to provide a bias current for Q3/Q5. It wont work as drawn.
Some amps can be a bit unpredictable if the input is made to float (even taking into account the resistor that isn't shown on the diagram) and I would hate for the amp to self destruct through untamed instability. That's why I recommended taking the input cap to ground.
Some amps can be a bit unpredictable if the input is made to float (even taking into account the resistor that isn't shown on the diagram) and I would hate for the amp to self destruct through untamed instability. That's why I recommended taking the input cap to ground.
Lifted the PCB from the housing and followed the traces, but could not find a resistor going from R2 to ground. Also measuring it with a multimeter does not show any resistance between ground and either side of that resistor (well, not "no resistance", but real real real high I presume 😉)
Disconnecting the signal line between preamp and poweramp, just before C1, reduces the noise slightly. It is still there, but lower in volume.
Disconnecting the signal line between preamp and poweramp, just before C1, reduces the noise slightly. It is still there, but lower in volume.
Strange.
The resistor would probably be around 150k (150,000) ohms to match R11. Can't explain that without seeing one for real 🙂
So the noise is still present. 150k for the feedback resistor (R11) does give a noise penalty, normally in modern design we would go for something in the 10 to 20k region to minimise thermal noise that all resistors generate.
The gain is set by the ratio of R11 to R2 PLUS 1 and so it is actually very high for a power amp, numerically 101. That's high gain, and high gain does mean noise. Typical gain values for a power amp would be around 20 to 25 with perhaps older gear at around 30 to 40.
So that's half your problem.
The preamp is also generating noise, which even though it may not be much is getting amplified by that massive gain figure.
The resistor would probably be around 150k (150,000) ohms to match R11. Can't explain that without seeing one for real 🙂
So the noise is still present. 150k for the feedback resistor (R11) does give a noise penalty, normally in modern design we would go for something in the 10 to 20k region to minimise thermal noise that all resistors generate.
The gain is set by the ratio of R11 to R2 PLUS 1 and so it is actually very high for a power amp, numerically 101. That's high gain, and high gain does mean noise. Typical gain values for a power amp would be around 20 to 25 with perhaps older gear at around 30 to 40.
So that's half your problem.
The preamp is also generating noise, which even though it may not be much is getting amplified by that massive gain figure.
I've had this amp too and I also had that hissing problem. My theory is that it is simply the nature of the beast and that it is due to poor quality design and components. It really did leave me with a horrible impression, as far as build quality would go. Looks like the kind of stuff you see in those modern fake old wood radios you see for sale in shady markets. Literally the cheapest/lowest quality amp I have ever seen. It also sounded really bad - normally an amp has clear sound and firm bass, but this had neither. It also picked up a lot of weird noise from my PC, which better amps don't do.
I think they themselves, due to not being good at their job, couldn't get rid of the hiss, so they cheated - if you turn the volume all the way down, it actually cuts off the speakers with a relay - to hide the obvious hiss which you would still hear with the volume all the way down.
It's a crap amp and not one worth playing music on.
I think they themselves, due to not being good at their job, couldn't get rid of the hiss, so they cheated - if you turn the volume all the way down, it actually cuts off the speakers with a relay - to hide the obvious hiss which you would still hear with the volume all the way down.
It's a crap amp and not one worth playing music on.
Last edited:
Unfortunately as others have said the hiss is inherent in the design . Tangent try really hard to give the impression of " serious " hi -fi but the reality is simply they're cheaply built down to a price and not to a quality. To be honest it could be worse the later amp cxeo model is a lot worse ( class D )
The PT2314 volume control IC has a noise level around -95 dBV over the audio band (see http://www.coldtears-electronics.com/images/PT2314.pdf ). After that you have an amplifier with an ideal gain of 101 times, as Mooly explained. That means you get about -55 dBV at the amplifier output. -55 dBV into an 8 ohm loudspeaker with a sensitivity of 90 dB, 1 W, 1 m would give you a noise SPL of about 26 dB at 1 metre distance in free space. Taking into account that you probably have two channels and assuming a room with a 1 metre reverberation distance, you could end up with 29 dB all over the room.
So I think it's a combination of the volume control IC's noise and the huge gain of the main amplifier. Reducing the ratio of R11 to R2 and changing the frequency compensation if needed seems like the logical solution. With a ratio of 24:1 instead of 100:1 you can still drive the amplifier to full volume with an input signal of 1 V, but you get about 12 dB less hiss.
Like Mooly, I also don't understand how they fixed the bias point without a resistor to ground after the DC blocking cap.
With the input of the main amp open rather than AC shorted, you eliminate the noise of the PT2314 but you aggravate the noise of the main amp, because its equivalent input noise current no longer gets shorted by the low output impedance of the PT2314. Maybe that's the reason why you have only little improvement with C1 disconnected.
So I think it's a combination of the volume control IC's noise and the huge gain of the main amplifier. Reducing the ratio of R11 to R2 and changing the frequency compensation if needed seems like the logical solution. With a ratio of 24:1 instead of 100:1 you can still drive the amplifier to full volume with an input signal of 1 V, but you get about 12 dB less hiss.
Like Mooly, I also don't understand how they fixed the bias point without a resistor to ground after the DC blocking cap.
With the input of the main amp open rather than AC shorted, you eliminate the noise of the PT2314 but you aggravate the noise of the main amp, because its equivalent input noise current no longer gets shorted by the low output impedance of the PT2314. Maybe that's the reason why you have only little improvement with C1 disconnected.
Last edited:
Measured again, and apparently I had my multimeter set on too low of a scale, but between R2 and ground there is a resistance of around 33K. So you are right that somewhere in the circuit that connection is there 🙂
For everyone else: thank you for taking the time to take a look at this, and for giving your honest opinion!
I have decided to not go tinker with this amp, as the results would turn out to be 'reasonable' at best. Instead, it is better off finding a new owner as-is while I find a different one, probably from a kit around here somewhere 🙂
Or it might just be sacrificed for parts
Its up to you 😀
Apart from the hiss you said you like the amp a lot. If it comes down to sacrificing the amp for parts, then why not try Marcels suggestion of pulling the gain down which can be done with just a resistor change.
Apart from the hiss you said you like the amp a lot. If it comes down to sacrificing the amp for parts, then why not try Marcels suggestion of pulling the gain down which can be done with just a resistor change.
Clocking noise on the I2C controllers. Even though the bus may be running in the MHz, the RF energy can spoil with low volume circuitry. (Analog Devices has a couple application notes on the topic.) The Texas Instruments PGAxxxx devices can sometimes have noise issues. They use a PTI controller which looks similar to the TI.
and the LMxxx voltage regulators are very noisy compared to discrete designs, or the newly introduced regulators from TI and Analog Devices
and the LMxxx voltage regulators are very noisy compared to discrete designs, or the newly introduced regulators from TI and Analog Devices
If it were my amp, I'd try something like the attachment: feedback resistor reduced to 33 kohm and 180 ohm emitter degeneration resistors added to the first stage to keep the loop gain and slew rate at essentially the same values as before the modification, while reducing sub-slewing TIM.
Of course I would also do some sanity checks before connecting expensive loudspeakers, such as:
-measuring the output DC voltage (should be at most a few hundred millivolts)
-checking that the output DC voltage doesn't change much when you touch nodes with an insulated miniature screwdriver (making very sure not to short anything)
-functionally testing it with a cheap loudspeaker
By the way, I doubt if the polarity of Q11 and the value of R? in the schematic are correct.
Of course I would also do some sanity checks before connecting expensive loudspeakers, such as:
-measuring the output DC voltage (should be at most a few hundred millivolts)
-checking that the output DC voltage doesn't change much when you touch nodes with an insulated miniature screwdriver (making very sure not to short anything)
-functionally testing it with a cheap loudspeaker
By the way, I doubt if the polarity of Q11 and the value of R? in the schematic are correct.
Attachments
Last edited:
I don't understand why the main amp stage gain was originally set to 100x.
Its almost certain that this is the main cause of the problem.
Typically you need 20x to 35x on a power amplifier - the higher gains for higher outputs where you want more swing.
I think the mods proposed by Marcel are a good step in the right direction.
I note that the tone control response leaves a bit to be desired - no RTZ for the treble - if you are p[icking up any HF noise before the preamp, that will just get blasted through if the treble is turned up. Bass response if also very idiosyncratic - peaks at 100 Hz and then drops off below that.
Its almost certain that this is the main cause of the problem.
Typically you need 20x to 35x on a power amplifier - the higher gains for higher outputs where you want more swing.
I think the mods proposed by Marcel are a good step in the right direction.
I note that the tone control response leaves a bit to be desired - no RTZ for the treble - if you are p[icking up any HF noise before the preamp, that will just get blasted through if the treble is turned up. Bass response if also very idiosyncratic - peaks at 100 Hz and then drops off below that.
Last edited:
That schematic is littered with errors. Q8 is supposed to be a 2N5551, that's not a pnp. The circuit seems inspired by a little Rotel of some description.
Possibly they were too lazy to use / set the built-in preamps (3.75 dB steps admittedly are a bit cumbersome), so they relied on the power amp for all the voltage gain. Too bad if you've got a noisy voltage control IC up front.
Really the only way of fixing this would be:
1. Drop power amp gain dramatically, as suggested.
2. Add another gain stage with a volume control (or attenuator switch) of its own up front. Maybe +20 dB with a 0/-20/-40 dB selection.
Possibly they were too lazy to use / set the built-in preamps (3.75 dB steps admittedly are a bit cumbersome), so they relied on the power amp for all the voltage gain. Too bad if you've got a noisy voltage control IC up front.
Really the only way of fixing this would be:
1. Drop power amp gain dramatically, as suggested.
2. Add another gain stage with a volume control (or attenuator switch) of its own up front. Maybe +20 dB with a 0/-20/-40 dB selection.
- Status
- Not open for further replies.