Problems intigrating headphone preamp with phono amp, fm transmitter...

[Bobsone]

Hi
First, I am a beginner, my skill level is about a 1.5 out of 10 and sorry for the long post, I wanted to include as much as possible.

I have been trying to find a 12 volt DC preamp to run some headphones and have run into a problem or two...
The preamp is to be mounted in a 60 watt Mitx case and be powered by the case´s DC 12V supply. The case already has a functioning phono preamp, FM transmitter and a soft latch switch circuit so the power supply can be switched from the front panel. The supply for the case comes from a 12V 60w DC brick. The build isn’t intended to be an audiophile quality system and the preamp will be used to supply some wireless headphones.
I have tried two different (cheap) Ebay headphone preamps (links 1 & 2).
1) Pure Class A Headphone Amplifier AMP 1969 Circuit DIY Kits für HD600 K701 | eBay
2) Portable Headphone Amplifier Board Kit AMP Module Kit For Classic 47 DIY +case | eBay

The preamp in link 1 works and doesn’t cause any problems when it is integrated into the Mitx case with the other components. Unfortunately the sound isn’t great, it is too tiring to listen to for any length of time. Based on this “test” I figured the inclusion of a preamp will work so I thought I would try a different option with opamps.

The link 2 opamp preamp also works well on the bench (powered by a SFX supply) and its sound is surprisingly good. Unfortunately it doesn’t work when it is connected to the other Mitx case components, it barely runs and produces lots of distortion. I discovered the problem is related to the input-output earths. I.e. the preamp doesn’t work when the earths are connected to anything in the case (e.g. phono amp output, case mounted 3.5mm headphone socket). But oddly it works well when the 3.5mm socket is removed from the case and only the RCA centre pins are connected.
I compared the two preamps and found what I think are a couple of issues with the link 2 preamp (I have tested two samples of the link 2 amp). First, it has unbalanced and (what I think is) high voltage outputs. I tested this by connecting the neg output to the multimeter and then moving the mm pos lead between the output pos L and R channels. The output volts are around 0.8V and 1.1V.
I then tried connecting the multimeter neg lead to the DC 12.2Vsupply neg terminal and found the signal input and output neg leads were both showing 6.1V.

I know the amps I have tried are not high quality but am unsure if the issues I am facing are the result of bad design, cheap components or something I haven’t thought of. Do opamp based preamps normally have a high voltage on the I/O neg leads...
I would like to have a preamp in the case but have run out of ideas-talent and I can’t find another 12V DC option that will fit.
So if anyone has a suggestion/an idea that might help or knows of a small(ish) form factor, good(ish) quality DC 12V preamp I could try I would greatly appreciate the input.


Regards.

 

[sgrossklass]

In #2 you probably shorted out the virtual ground (at half supply) to real ground. Looks like this may be a split supply circuit not 100% adapted to single supply operation, which would have to be done properly. There's some small but decisive differences. Without a schematic, some reverse-engineering will be required first.

#1 is derived from the original JLH speaker amp, not the headphone version. Interesting. No idea what kind of performance to expect here TBH. I suggest you check output midpoint voltage (in front of each output coupling cap) and adjust it to something sensible like around +6V with the multiturn trimmer if need be.
Hey, that doesn't actually look too bad in sim, assuming midpoint voltage is set up properly. I'm getting rather more output stage current than expected though, so I don't think the schematic given in the auction is entirely accurate, or maybe it's just a transistor model discrepancy.
Is the circuit getting nice and toasty as it should? If you fire up LTspice, you could compare voltages in sim to the real deal (sorry, I did not watch out for component designators, so you may have to correct those; the text file with the models can be had at Bob Cordell's website).

 

[Bobsone]

Hi
Thanks for the reply sgrossklass.


I have been playing (attempting to fix-improve) the two preamps. I thought I may have messed up with the assembly of kit #2. Because they are cheap I have now assembled two #2 kits and the both demonstrate the same voltages...
I quickly run out of talent with them (it looks like they are supposed to be like that), so I have been investigating the #1 kit.


Again I am starting from the view that I did something wrong and have gone over it as best as I can, I can’t see anything wrong but this was my first foray into surface mount components so perhaps one isn’t perfect, I will return to this one later... I have the VR-VL voltages set at exactly 50% of the vcc voltage.

Plan B, I have searched the net for a headphone preamp circuit diagram that I could have a go at assembling but haven’t found much, and I really am unsure how to discern the quality of the circuit. I thought the preamp aspect of this would have been the easiest, it doesn’t need to power anything like 600ohm headphones.

So, does anyone have/know of a 5 or 12volt DC headphone preamp circuit (that isn’t enormous or overly complicated) that I can have a go at assembling-test on a breadboard? The mITX case supply also has a -12volt supply but I doubt that will have much current. It only needs a volume control i.e. it doesn’t need external tone, base or balance controls.

Thanks.

 

[sgrossklass]

I have been playing (attempting to fix-improve) the two preamps. I thought I may have messed up with the assembly of kit #2. Because they are cheap I have now assembled two #2 kits and the both demonstrate the same voltages...
I quickly run out of talent with them (it looks like they are supposed to be like that), so I have been investigating the #1 kit.

I am not quite sure where the problem with #2 is though. It looks like it's a take on the classic apheared-47 circuit, adapted for single supply with a cMoy-like virtual ground. So my educated guess at the schematic would be something like this:

(Vplus / Vminus = power supply connection, so Vminus is your "real" ground, while the ground symbol is attached to the virtual ground; RL = external load)
(EDIT: Sorry, I forgot the small film cap of ~100 nF, which should go across Vplus-Vminus at U2.)
Which has some issues like bad PSRR but I have no reason to believe that it should not work, unless there were something stupid like a routing mistake or reversed electrolytic caps going on. (With these cheap Chinese kits, you have to be prepared for goofs like that.) Please verify that output jack ground goes to Vminus rather than virtual ground. It is easy to misroute when the layout software is a fancified drawing program with no physical checks like EAGLE.

Are the boards you got the exact same ones as shown in the auction? I have seen some other variations suitable for split supply only.

Incidentally, I'd suggest a slight modification to the virtual ground section in order to fix the PSRR problem:

IOW, remove the capacitor that goes from Vplus to virtual ground and reroute its negative leg to Vminus instead.

Even assuming you do get things working, however, there is one major problem you would be very likely to run into: A first-rate ground loop when using both a PC's internal power supply and its audio output.
A lot of people have run into this with their DIY boomboxes when they would be installing both an audio amp board and either a Bluetooth receiver or Raspberry Pi or similar and then attempt to run both from the same power supply - they'd be greeted with annoying ground loop noises. (This is because part of digital source power supply return current would now be flowing over audio connection shield.)
Fixing this usually involves either a line-level isolation transformer for the signal or using an extra (galvanically isolated) power supply, though basically it can be done without these. Grounding in audio is tricky business!

I thought the preamp aspect of this would have been the easiest, it doesn’t need to power anything like 600ohm headphones.

What should it be able to drive then?

Contrary to popular belief, 600 ohm (and other high impedance) loads are not actually "hard to drive", quite the contrary in fact. They're easiest on the output stage and all, and merely become inconvenient when you just cannot supply them with enough voltage for a decent amount of output, i.e. low supply voltages. Reducing load impedance addresses that, but it inevitably makes the output stage work harder. Less voltage, more current. Headphone efficiency is another major factor, as it may vary quite significantly depending on construction. Planar magnetic headphones are particularly notorious for being inefficient, while IEMs with balanced armature drivers can be extremely sensitive.

 

[Bobsone]

Hi
Thanks again for the reply sgrossklass.

First, some numbers;
Using the schematics posted by sgrossklass as a reference, R3, R4 are 2.7K and C4, C5 are 470uF 25v electrolytic caps. The psu is delivering 12.1volts, the voltage between R3 and R4 is 6.07v and the voltage on the grounds of the input and output is also 6.07v @ 4.5ma.

Please verify that output jack ground goes to Vminus rather than virtual ground.

As far as I can tell, the input and output grounds return to the virtual ground between R3 and R4. I have drawn this conclusion because I am seeing a 2.7K load when I test between the input and/or output grounds and V-. Also, when I test between the input/output grounds and the output of R3 (input of R4) there is 0 ohms.

Are the boards you got the exact same ones as shown in the auction? I have seen some other variations suitable for split supply only.

I can confirm that both amps are the same and they both work the identically, I have confirmed the correct orientation of all electrolytic caps.

Incidentally, I'd suggest a slight modification to the virtual ground section in order to fix the PSRR problem:

The virtual ground layout looks like the one in the first link posted by sgrossklass with 12.1 volts going to C4 and R3 and 6.07V (from R3) going to the C4-C5 junction and to R4.
I have tried the alternative design where C4 and C5 have separate connections to V-, unfortunately this test didn’t alter anything.

Even assuming you do get things working, however, there is one major problem you would be very likely to run into: A first-rate ground loop when using both a PC's internal power supply and its audio output.

Sorry, I wasn’t specific in my first post, I am using the ITX case as a housing and power supply for a phono amp, a headphone preamp (hopefully) a soft latch switch and an FM transmitter, there wont be any PC components in the case. So far the set up is working well, now I just need to solve the headphone preamp issue .

So, would diverting the input and output grounds to V- cause any problems?

Thanks for your time
B

 

[sgrossklass]

As far as I can tell, the input and output grounds return to the virtual ground between R3 and R4.

There's your problem.

I can only guess that someone wanted to convert a split supply circuit to single supply (else why would there be coupling caps?) but simply forgot rerouting input and output ground. Maybe they did eventually notice but not until the boards had already been made. So what do you do? Oh well, just dump 'em on eBay. That's cheap Chinese kits for ya...

I have tried the alternative design where C4 and C5 have separate connections to V-, unfortunately this test didn’t alter anything.

Better do this later, ideally with a noisy power supply so you can appreciate the difference.

So, would diverting the input and output grounds to V- cause any problems?

Not if you do it correctly (obey star grounding and all).