Gain and grounding: have I got it right?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Everyone,

I've spent a lot of time reading this forum, and I think I understand how to apply the ideas here to my scenario. But if you would read over my approach and critique, I'd really appreciate it:

My sister has an electric cello, the internals of which consist of a pickup and 1/4" jack-- it doesn't contain any active amplification. I am building her a combination practice amplifier and pre-amplifier for when she has gigs. I'm using the TDA7052A chip IC which is simple and sounds nice.

I don't want to screw around with mains voltage at all, so I am using a 9V wall wart with a jack onto it which will plug into the enclosure. Then I'm stepping the voltage down to the chips' ideal 6V using a 7806 regulator. I'm building it into a conducting enclosure (metal box).

So for the grounding, I should star-ground everything internally, but not ground the chassis to the signal ground, right? There's no mains voltage inside the chassis so there's no safety concern, and the signal ground isn't connected to earth ground, so it wouldn't help shield the internals from RF, is my reasoning.

Also, as far as gain, it's variable with a pot up to 40x. That's about the right range to take a cello pickup up to line level, right? The chip will drive a small speaker from a tape head, so I'm thinking it would be about right for this as well.

Lastly, anyone have a clever method of heatsinking a dip-8 package? This little guy gets hot at the full 1W output.

Suggestions, critiques, rude gestures all appreciated!
 
I don't want to screw around with mains voltage at all, so I am using a 9V wall wart with a jack onto it which will plug into the enclosure. Then I'm stepping the voltage down to the chips' ideal 6V using a 7806 regulator.

Why bother? The TDA can work from as much as 18 V.

So for the grounding, I should star-ground everything internally, but not ground the chassis to the signal ground, right?

Rephrase that to, you don't need to ground the chassis, but it helps. Such a connection adds virtual mass to the ground and makes it more stable.

Also, as far as gain, it's variable with a pot up to 40x. That's about the right range to take a cello pickup up to line level, right?

That is a question you'd better put in the Musical Instruments section of the Forum.

Lastly, anyone have a clever method of heatsinking a dip-8 package?

How about self-adhesive heatsinks? They are widely available for computer RAMS, but also for DIL packages.http://www.conrad.de/ce/ProductDeta...99/8700/8700/8706/870611_LB_00_FB.EPS_400.jpg
 
1 Watt to amplifiy a cello at a gig????? Unless you use a very sensitive speaker im guessing the unamplified sound from the cello will be louder than your 1 watt with the tiny speaker you propose.

Hi Mark,
It's an electric cello. There's a pickup inside like an electric guitar, but there's no sounding board, so the cello doesn't make any noise without an electric amplifier.

Why bother? The TDA can work from as much as 18 V.

Hi PacificBlue,
That was my initial thought since that's what the datasheet says, but I've found in practice that it really wants 6V-- anything above that and it gets really, really hot. Using the 6V regulator has the additional advantage of removing ripple current from my cheap power supply, so there's that.

You don't need to ground the chassis, but it helps. Such a connection adds virtual mass to the ground and makes it more stable.

It seems like since it's not connected to earth ground, any voltage the chassis came in contact with would affect the signal, rather than being drained to earth, or am I wrong?

After thinking about the gain issue, the maximum gain of the chip is set to 40x, so if I end up needing more than that I'm sunk anyway. But, if I put a 10K pot on the incoming signal, she can trim the incoming signal to a lower range if she needs it.

Thanks much for your informative reply!
 
It seems like since it's not connected to earth ground, any voltage the chassis came in contact with would affect the signal, rather than being drained to earth, or am I wrong?

It is actually the other way round. Since it has no contact to earth, there should be no interaction. Pretty much like birds that can sit on high-tension power lines without being affected, because they do not come into contact with earth potential at the same time.
 
HI
I assumed it was an electro acoustic.
I would suggest an electric guitar preamp/ buffer as pickups can be affected by the wrong input impedance. Then connect output of preamp/ buffer to small amplifier. This will allow you a much wider choice of more suitable amplifier chips.
Mark
 
40x gain might be out of range for that chip, and causing oscillation...hence why the chip is getting so hot. You might also find that that 9v wallwart, if it's unregulated, is producing a much higher voltage and thats why the chip is hot, although I wouldnt expect it to produce more than 14v even off load.
 
HI
I assumed it was an electro acoustic.
I would suggest an electric guitar preamp/ buffer as pickups can be affected by the wrong input impedance. Then connect output of preamp/ buffer to small amplifier. This will allow you a much wider choice of more suitable amplifier chips.
Mark

Hi Mark,
I've been reading up on this, and it's sounding like I might want something more like a traditional guitar preamp circuit. Which makes sense, actually. I'm going for something that sounds decent and is pretty rugged-- that's what attracted me to the TDA7052A.

So based on that criteria, do you really think I would benefit from a separate let's say FET based pre-amp? Or should I say good enough with what I've got and start building the circuit?

Thanks again for the help!
 
Okay, having bought some parts and read a lot more about this, the signal path I'm looking at is:

1. a low pass filter, with the corner frequency set at about the limit of human hearing.
2. a discrete FET buffer amp, with a 3Mohm impedance input, to handle the piezo pickups.
3. The output jack, so connected so that plugging a cable into the jack disconnects the speaker. Otherwise, it's in practice amp mode, so the signal is passed to:
4. The TDA7052A amp and 8ohm speaker.

Attached is a schematic of the above as I'm planning to do it. I tried to breadboard it but with all the wires everywhere it's more of a radio antenna than anything right now. I'm going to try making the pcb and wiring one up, let you know how it goes.
 

Attachments

  • Practice Amp rev1.png
    Practice Amp rev1.png
    13.7 KB · Views: 112
You have two filters at the input
HiPass 1uF & 3M F-3dB ~0.05Hz.
LoPass 2n2F & 4k7 F-3dB ~15kHz.
Both are ridiculous.
for normal hearing we usually set the passband to 20Hz to 20kHz. However, many use 0.5Hz to 10Hz as a HiPass and 100kHz to 300kHz as the LoPass, that let's through all the Audio in the signal.
Your guitar can manage with a more restricted passband, maybe try 10Hz to 50kHz.
Any additional shaping of the "music" should be done in some kind of adjustable pre-amp designed for Guitar.
 
The high-pass is indeed somewhat exaggerated, especially as there is a second high-pass after the FET formed by 10µF and the amplifier's input impedance 20k ±5k which leads to already extremely low 0,63-1 Hz.
According to this link there is no significant output from a Cello above 8 kHz, so 15 kHz low-pass should be acceptable. On the other hand there is no recommendation for a low-pass filter in the TDA7052's datasheet, so you could probably just skip it.

If that FET-buffer is supposed to be J. Donald Tillman's, there are two resistors missing.

You should increase the power supply capacitance. both before and after the voltage regulator and add the 100 nF decoupling cap close to the TDA's supply pins as specified in the datasheet.
 
Thanks once again to both of you for your insightful comments. The design I was working off was based as much on this one:
Basic Buffers
as J Donald Tillmann's design. The resistors that were left out were because I couldn't understand what they did, and they were not included in the "basic" jfet buffer shown in the link. I'm going to include spots for them in my PCB and then I can experiment on the real device instead of the breadboard.

Revised schematic attached. I'll let you know how it goes.
 

Attachments

  • amp rev2.png
    amp rev2.png
    15.6 KB · Views: 51
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.