Hi guys,

I was wondering if you could help me with some problems I'm having with a home project? The short (ish) story is that I made myself a DIY 5.1 channel amplifier for use in a home theatre set up. However, when amp is connected to an HTPC's analogue out's and makes a bit of a thump when it (the PC) turns on (I wonder that there is a DC offset to the sound card, however it's a bit of a high end card so I don't want to mess with it).

As this project represents my first foray into home electronics since my A. Levels (almost two decades ago), I turned to the internet for answers and figured that I'd put together Project 38 from the ESP site. However, rather than having the relay switch the amplifiers power supply (which is already switched by the computer's power supply), I would use a N/C relay which would trigger the kit board's sleep/mute function when there is no signal and break that loop when there is an input signal.

I can't seem to get the circuit to work on a prototype board (though instead of a relay I am using an LED, so perhaps that is incorrect?). Initially, it did seem to work correctly when I applied a signal to it, but I'd been testing it using an iPhone as a source, and it seems that it was actually turning on when I touched the iPhone rather than when the music was started.

I've checked the layout of my prototype circuit through a few times and I'm pretty certain that I have it right, I was wondering though if anyone could help with pointers to figure out what I've done wrong. Testing equipment wise I have a digital multi meter, but I understand with OP-AMP circuits I can't really use this as it is likely to change the input/output gain on the OP-AMP perhaps affecting the circuit itself.

Thanks in advance,
~Dan

 

Welcome to AAC!

You will be happy to find you have continued a long standing tradition here at AAC. Without schematics or links (which should have schematics) we can't really help. You can post a graphic image using the <Go Advanced>/<Manage Attachments> buttons on the bottom of the post. Preferred file formats are .png or .gif .

I mentioned this because it has become a source of humor among us, you will find this to be an extremely friendly group, and a flame free zone.

****************************

Opps, missed the link. My bad.

What is your power supply source, that is where I would start first. If it is a battery(s) then we will move on down the line into the circuitry.

This is the circuit we are talking about, correct?

 

1) Measure your input voltage from the junction of RL1/D2 to ground. You should get the same voltage on both sides of the relay if Q1 is not conducting
2) Measure across C3; you should get 1/2 the voltage you measured in step 1.
3) Measure the voltage from U1 pin 3 to ground; you should measure the same as step 2.
4) Measure from U1 pin 5 to ground. You should measure roughly 91% of your reading in step 2.

 

Hi guys,

Thanks for your replies. Bill, no worries about the link, I should have made that a bit clearer ... though for a moment I thought the link had been removed by moderation. You are correct though, that is the circuit I am trying to build.

To answer your questions, with the prototype board I am using a cheap lab power supply that i got from a local electronics store. I've set a voltage of 12V and a max Amperage of 0.2A. In the actual build itself I intend to take power from the computer's 12V rail (which I believe can deliver up to 4A). The supply sets itself to constant voltage when on and the circuit draws around 0.01 amps (the amperage draw is from memory, I can't check exactly right now).

SgtWookie, I shall take some readings later tonight and post those when I have them.

Thanks again,
~Dan

 

You will note R8 and R10 makes a simple voltage divider, the voltage will be approximately ½Vcc, or 6V. It is loaded here and there with other resistors, so I could be off a little.

It biases the other op amps. This means that voltage will be recreated throughout the circuit, by the negative gain of the op amps.

I am not too fond of R9/D1 feeding Q1, this means Q1 is partly on all the time, and the relay is always getting current (just not enough to turn on). This could be very related to your problem, I think the relay is being used to translate analog to digital.

The internet is full of circuits that are fundamentally flawed. This may be one.

 

Hi guys,

Just a quick update. I've had an evening of technology meltdowns with successively the comms board on my iPhone dying turning it into an iBrick, my laptop hard disk having a MBR episode and not being able to boot and one of the disk's in my NAS server packing up . Having fixed the laptop and managing to extract a valid backup from the phone, I think that it's time I called it a day on technology (for today at least).

Sorry though as I've not had time to look at taking those measurements, nor digest what you commented on last Bill ... though something in the back of my brain is prompting me with something I read in your LEDs, 555s and PWM eBook about cleanly switching transistors and heat ... I can't seem to find the reference quickly, but is that what you're talking about or is it that the relay is receiving current which is bad?

Hopefully when I get home from work tomorrow, have fed, bathed and put my daughter to bed, eaten myself and can sit down for the 20 minutes of the day that I get for hobbies ... I'll get a chance

Thanks for your help so far guys.

~Dan

 

Right, I managed to get some measurements. I'm hoping that I've done it right; and yes, I am away that I have thrown myself into the deep end of understanding here, so any pointers to online reading would be really appreciated.

A few observations from measuring the bread board build;

• The lab power supply seemed to be putting out around 11.84V whilst it's display stated 12V. I left it at this setting.
• I replaced C4 with a 1uF cap, this meant that it drained faster so when the circuit powered on I didn't have to wait 20 minutes for it to deactivate.
• As stated before, I replaced the relay with a orange LED and 510Ω resistor. The LED came from a lucky dip bag and lists its forward voltage at somewhere between 2.0~2.2V.
• Throughout the test, I left the circuit unconnected and removed the 3.5mm jack wire. The circuit seemed to activate on its own, sometimes briefly, other times for longer. I couldn't attribute it to anything specific, though wonder whether it is picking up RF from the wireless in my laptop.
• I wonder if this could be a ground issue, my lab power supply only came with 2 connectors, so I am having to fashion a ground by jamming a wire in the connector and trusting that it has a good connection.

Measurements were as follows:

1. The input voltage measured 10.26V at the RL1/D2 junction when the circuit was inactive. The opposite site measured 11.83V. I noted when active the voltage was 0.01/1.91.
2. 5.63V
3. 5.12V - no matter how careful I was here in ensuring that I wasn't touching any other component, as soon as I touched pin 3 the circuit activated.
4. 5.07V (I make 5.12V as 91% of 5.63V)

Bill, I've still not had a chance to fully digest your last post. Fundamentally though, yes I believe that it is trying to turn the analogue signal of the music source into a digital on/off for the relay.



I would appreciate any suggestions or pointers for alternatives to consider. Technically I don't think that I have the expertise to design something like this myself, though I am comfortable building from a schematic (hence picking up this project). If though the design is flawed, then I guess I will have to work a little harder on this and would really appreciate any help I can get.


Thanks again,
~Dan

 

Something is causing U1 pin 3 to be pulled down a bit. It might be that R3 is open, C1 is shorted, pin 3 is shorted to pin 2, or something of that nature.

Try measuring from U1 pin 3 to the junction of R1/R2 (across C1). If the voltage is zero, C1 could be shorted.

 

Hey guys,

Sgt. Wookie, I think that you were right and there must have been a short somewhere. As the prototype was on bread board I ripped all the parts out and laid everything out again without trying to take any short cuts instead of trying to track it down. I've ended up using all of a 64x20 bread board, and far more jump wires than I think is necessary, but it works! The circuit now only activates when I play music.

I'm going to go to some vero board from here and see if that achieves the desired objective in-situ.

I've been doing some reading in an attempt to understand what the design is doing; if I understand what Wikipedia is telling me U1A, R5 and R4 form a negative feedback loop with a set gain (the 6V plus whatever the music signal adds is being amplified) and R6, R7 & U1B form a comparator which amplifies the difference between the signal source and the ~6V? I think that given that there would always be a difference to be amplified that is why you are saying Q1 is always active Bill? Forgive me if it is a ridiculous suggestion but would something like a zener diode work to prevent that; as I understand them the will not allow current to flow until a specific voltage is reached, approx 3V which is what is needed to turn on Q1? Is the nearly on state not good for the mosfet (i.e. will lead to it running hot all the time)?

Otherwise though, from what I am seeing the circuit I have appears to work. I'm interested to hear any suggestions on design revisions.

Thanks in advance.
~Dan

 

I am not too fond of R9/D1 feeding Q1, this means Q1 is partly on all the time, and the relay is always getting current (just not enough to turn on). This could be very related to your problem, I think the relay is being used to translate analog to digital.

I'm afraid you're mistaken here.
U1B is being used as a comparator; the output either high or low.
It's detecting if there is audio output from U1A, with the threshold being set by R6/R7. If U1B's output is high, C4 and Q1's gate is charged via R9 and D1; R9 to limit the maximum current flow, and D1 to keep C4 and the gate charged when the output returns low.

The internet is full of circuits that are fundamentally flawed. This may be one.

That's a Rod Elliot design. I haven't seen anything on his site that's fundamentally flawed. There are a few "loose ends" here and there, but it's really quite above the ordinary.

 

I've been doing some reading in an attempt to understand what the design is doing; if I understand what Wikipedia is telling me U1A, R5 and R4 form a negative feedback loop with a set gain (the 6V plus whatever the music signal adds is being amplified)

The 6v is simply a "virtual ground" reference point for the amplifiers; as there is no negative supply. This has to be done with single-supply amplifiers, or else the negative portion of the input signal would be truncated.

The input signal is amplified by -R5/R4, or -100 times.

and R6, R7 & U1B form a comparator which amplifies the difference between the signal source and the ~6V?

U1B is used as a comparator. This is all explained in Rod Elliott's notes.
R6/R7 establish a reference threshold. If the input signal from U1A exceeds the threshold, then the output of U1B goes high, and Q1 is turned ON.

I think that given that there would always be a difference to be amplified that is why you are saying Q1 is always active Bill? Forgive me if it is a ridiculous suggestion but would something like a zener diode work to prevent that; as I understand them the will not allow current to flow until a specific voltage is reached, approx 3V which is what is needed to turn on Q1? Is the nearly on state not good for the mosfet (i.e. will lead to it running hot all the time)?

It's really not necessary. Current flow through the relay coil should be pretty low anyway; perhaps 30mA-80mA. At worst case, Q1 would dissipate 1/2 the power that the coil would when it had full voltage across the coil. The amount of heat would be trivial.

If one really wanted to, they could monitor the voltage on C4 using a Schmitt-trigger to turn the MOSFET on and off; but that's adding more parts to the mix.

 

Great,

Thanks Sgt. Wookie for your help and your explanation. I think then that I'll not modify the circuit beyond what I originally envisaged, which was to replace R6 with a variable resistor as Rod suggests in his notes. This way I'll be able to tune the sensitivity once the circuit is in place.

As an aside, I've been looking at producing a PCB schematic in eagle so I can use this design on other projects. The auto router has been giving me issues as it seems to have swapped components around. But that is a topic for another post if I can't figure out what is happening myself.

Thanks again for your help.
~Dan

 

Wookie, take another look at this circuit. The signal is pulses matching the spikes of the signal, which is rectified and fed into the MOSFET as an analog signal. The comparator does not stay on or off, but switches in time with the music.

It is not a clean switch either, as the 1458 is a 741 under a different name.

I do not have an easy answer, and if it works good enough...