| Author |
 Topic  |
|
|
pawnshop_trash
Gold Member
USA
603 Posts |
 Posted - 10/22/2005 : 00:12:10
|
alas, this isn't for a Boss repair or mod project, or for an audio signal at all... I'm trying to amplify a variable (~15 mV max) analog signal from a load cell, to get near 10 V DC. a simple non-inverting op amp should do the trick, but what op amp brands/models are the best if low noise is important? (e.g., from another thread, bossarea said that the JRC 4558D is single shielded, while the 4558DD is double shielded, for lower noise; do all op amps have the same shielding nomenclature?)
likewise, what type of resistor is best for low noise? (e.g., carbon film, ceramic, metal film, metal oxide film, etc?) while quality is an issue, it would also be great if I can make a prototype this weekend from cheap parts bought at Radio Shack, as a 'proof of concept' if nothing else. then all I would need to do is order higher quality parts from DigiKey or someplace. any and all recommendations will be greatly appreciated!
oh yeah, please also feel free to post the brands/types of resistors or op amps (or capacitors) that would be best for audio electronics applications, especially Boss pedal mods. |
Edited by - pawnshop_trash on 10/22/2005 00:21:36 |
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/22/2005 : 22:52:13
|
| You want to amplify an AC signal in order to get a flat DC signal? That doesn't even make sense. |
 |
|
|
arcanon1313
Silver Member
USA
414 Posts |
Posted - 10/22/2005 : 23:38:55
|
| Going from what Robert Keeley says the metal film resistor is lower noise then all other ones. But I maybe wrong. |
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/23/2005 : 05:20:15
|
quote:
Originally posted by arcanon1313
Going from what Robert Keeley says the metal film resistor is lower noise then all other ones. But I maybe wrong.
I don't know how metal oxide resistors stand, but according to the book that I used for my analog electronics class, wirewoud resistors have the lowest noise, followed by metal film. I think metal oxide resistors came out after this book was written. I've never heard of ceramic resistors (perhaps you're thinking capacitors?). And the lower noise that you experience is lower flicker noise, Johnson noise is the same no matter what material the resistor is made of. You also have to worry about shot noise, which is dependent on how much or how little current is passing, and has no bearing on what the resistor is made of either.
I've only seen wirewound resistors as high-power low-resistance devices. The same is basically true for metal oxide resistors. I'm guessing that they are probably not for your application.
I don't think any manufacturer of resistors is better than the other. As long as they have the same ratings and materials there is nothing to differentiate them.
I still pose my original question, which I think is the more important one.
quote:
Originally posted by walrus121
You want to amplify an AC signal in order to get a flat DC signal? That doesn't even make sense.
|
|
|
|
pawnshop_trash
Gold Member
USA
603 Posts |
Posted - 10/23/2005 : 06:31:45
|
walrus and arcanon, thanks for the info!  I had checked out DigiKey.com, but was just bewildered by the number of material choices available for resistors. to obtain the required gain (~650), 65k Ohm and 100 Ohm resistors (or a similar combination thereof, e.g. 650 Ohms and 1 Ohm) will work for an op amp. I'll look for 1/4 watt (or better) wire wound or metal film resistors in appropriate sizes.
incidentally, the signal to be amplified is DC. typically, a transducer like a load cell requires a constant voltage (here, 10 V DC) as an "excitation" voltage, and the output of the load cell is a smaller voltage that varies with (and is proportional to) the amount of force placed on it. simply using a multimeter to read voltage is fine for incremental step force measurements, but the loading curve that will eventually be used (and must first be verified with this load cell) is very fast and nonlinear. thus, some sort of computer DAQ interface is needed to capture a few cycles of the loading curve, and since our USB-based DAQ device doesn't have enough on-board amplification available, an op-amp is required. |
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/24/2005 : 04:33:14
|
| I think what you are reffering to is known as an AC signal with a DC offset (sum of an AC and a DC signal). |
|
|
|
bossarea
Forum Admin
United Kingdom
3652 Posts |
Posted - 10/24/2005 : 09:57:49
|
Sounds like pure DC amplification to me.
Why do you need the high spec components. No one is gong to listen to the output of this opamp so as long as the DC voltage is detected as properly, the component quality shouldn't matter that much. Am I missing something? If speed is of the essence you may want to look at opamps with high slew rate values too. |
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/26/2005 : 07:16:37
|
quote:
Originally posted by bossarea
Sounds like pure DC amplification to me.
For the guy who has completed every electronics course that the Department of Physics offers plus his independent studies, who is doing his honors thesis on an electronics topic, who plans to go to graduate school in electrical and computer engineering, please define what "DC amplification" means.
quote:
If speed is of the essence you may want to look at opamps with high slew rate values too.
Transistors still have op amps beat. Although at that point we would be talking about low-frequency radio oscillations. Often op amps with higher slew rates cost less than the old school boys and have the same pinout.
Back to resistors, something that I forgot to mention: wirewound and edgewound (basically a subtype of wirewound) resistors are also inductive. They are basically are low/moderate inductors with high DC resistances. Therefore they aren't really suitable for AC applications. In that case you would use metal oxide resistors. |
|
|
|
bossarea
Forum Admin
United Kingdom
3652 Posts |
Posted - 10/26/2005 : 13:48:03
|
quote:
Originally posted by walrus121
please define what "DC amplification" means.
If you connect an opamp in a standard non-inverting configuration (R1 between minus and ground, R2 between output and minus), any DC voltage applied to the positive input of the opamp will be amplified to the output by a factor Av = 1 + R2/R1.
This is commonly used where a sensor gives out a variable DC voltage between i.e. 0-20mV while the rest of the circuit expects an input voltage between 0 and 1V. We will then need to amplify the DC voltage 50 times to match the sensor to the circuit. In this example we could use R1 = 1k and R2 = 49k to achieve Av = 50.
At no time are there ever any AC voltages in this circuit so the amplification we have must be pure DC.
|
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/27/2005 : 08:12:59
|
Yeah, I decided to look in the index of my book and there are a few entries for DC amplification. It was never talked about in my class though. One of the entries just said "Ch. 7," it didn't even give a page number, that was real helpful (and no, Ch. 7 is not solely about DC amplifiers). The only application of a DC amplifier that it gave was with a really complicated example circuit that uses four op amps (not all of which are the same model), three transistors (also not all of which are the same model), two p-channel mosfets, two diodes, a DP3T switch and a lot of resistors. I didn't even really understand why the circuit would be useful even though they explained the application.
You call the input voltage "variable," so I guess that you mean that the voltage is basically DC but is subject to some sort of not-necessarily-predictable long-period fluctuation. I thought by variable you you meant AC. I guess that we are all clear now.
My advice on the resistors still stands. |
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/27/2005 : 08:47:36
|
quote:
Originally posted by pawnshop_trash
but what op amp brands/models are the best if low noise is important? (e.g., from another thread, bossarea said that the JRC 4558D is single shielded, while the 4558DD is double shielded, for lower noise; do all op amps have the same shielding nomenclature?)
Since you mentioned the JRC 4558D and and double-D, the extra shielding is of limited use if the box that you are wrapping your project in is already made of metal or has some other conductive substance on it. Niether version of the JRC 4558 is actually considered a low-noise op amp. Any op amp with the same pinout, a supply voltage range that contains your supply voltage and meets or exceeds the following characteristics is substitutable.
- Slew rate
- Equivalent input noise voltage (lower)
- Unity gain bandwidth
- Peak-to-Peak Output Voltage Swing (not necessarily relevant unless you are pushing the limit)
- Temperature operating range
As for the nomenclature, there is no standard, and every manufacturer does their own thing. The only exception to this is when one manufacturer copies another manufacturer's part.
For the JRC 4558D I reccomend substituting with an LM833N, but that is based on the conditions and demands of an effect pedal, and I don't know exactly what you are doing.
Capacitors depend on your application as well. Teflon capacitors are the best from what I have read, but I have never seen one. Second best are polycarbonate, polypropylene and polystyrene, and your choice depends on your application. Next comes glass, then porcelian (low value caps). Mylar/polyester film capacitors are good and are cheap. Ceramic capacitors are not good. Electrolytics and tantalums are the worst that you can use, but are available in really high values. Tantalums are better than electrolytics though. Double layers also suck and are available as the highest values but I have not seen one just like I have never seen a Teflon capacitor.
Anything else that you need? |
Edited by - walrus121 on 10/28/2005 21:55:16 |
|
|
|
pawnshop_trash
Gold Member
USA
603 Posts |
Posted - 10/27/2005 : 17:51:17
|
thanks for the tips, walrus! I took a couple of instrumentation/measurements classes in school (mech eng'r), but the most recent one was 9 years ago and I'm not sure if I ever actually built anything electronics-wise (rewiring a guitar with passive pickups isn't nearly in the same league).
for this 'proof of concept' non-inverting amplifier, I ended up using metal film resistors and the Radio Shack 741 op amp, and it worked (although the gain was about 3/5 of what it should have been theoretically). based on your advice, I'll look for higher quality components when I build it in a more permanent box.
bossarea, the main reason I wanted higher-quality/lower noise components is because the signal to be amplified was relatively weak (max ~15 mV) relative to the noise floor of the USB-based DAQ device (~1 mV).... |
Edited by - pawnshop_trash on 10/27/2005 17:53:01 |
|
|
|
walrus121
Silver Member
USA
187 Posts |
Posted - 10/27/2005 : 18:34:06
|
quote:
Originally posted by pawnshop_trash
for this 'proof of concept' non-inverting amplifier, I ended up using metal film resistors and the Radio Shack 741 op amp, and it worked (although the gain was about 3/5 of what it should have been theoretically). based on your advice, I'll look for higher quality components when I build it in a more permanent box.
The 741 op amp was the first op amp and is among the worst as far as noise is concerned, and has some other abyssmal characteristics. Probably the only reason why it is still on the market is because of sentimental value. If you want to make sure that the gain is right, look for an op amp that is "internally compensated for unity gain." I think that you will get better results that way. |
|
|
|
pawnshop_trash
Gold Member
USA
603 Posts |
Posted - 10/27/2005 : 19:54:23
|
| thanks again, walrus! I had mounted an 8-pin IC socket to the breadboard, so I should be able to change out the op amp to a better one later (provided each pin has the same function, I guess). you're right about the 741's noise; with the load cell not loaded, the op amp's output is about 0.766 V instead of near zero.... |
|
|
|
coxster
Copper Member
USA
3 Posts |
Posted - 11/02/2005 : 19:28:03
|
I think you're off to a great start. I personally don't believe that there's such a thing as a low noise op amp. When you compare 50 transistors to 2 or 3, it's obvious. i've always used TL series ( back when I designed stuff at Conklin ) and was never dissapointed.
are you using the input analog signal to simiply trigger a more 'digital' device???? If that's the case, i remember a PAIA circuit from 20 years ago that used a high gain op amp circuit, with diodes on the inputS, so the pos pulses were amplified by the non-inv input, and neg pulses were made pos on the inv input. But then again as low a signal as you have, you may never exceed the breakover voltage to forward bias the diodes
Good Luck
Jim Cox
playforhim@gmail.com |
|
|
|
pawnshop_trash
Gold Member
USA
603 Posts |
Posted - 11/06/2005 : 07:03:01
|
thanks for the tips, Jim! in my case, the input analog signal from the load cell (Interface SML-1000; see http://interfaceforce.com/mbsps_loadcells.html) is just too weak to be read (after A/D conversion) by the USB-based Data Acquistion (DAQ) device I have. so this op amp is not really a trigger, just an amplifier....
but after 5 circuit designs (the last two of which were designed on a breadboard), I *finally* got an op amp to work for my application. as a general rule, I hate using breadboards because of the inherent uncertainty in whether or not an electrical connection is really made (esp. since the ones at my job are old and pretty much crapped out). but after 3 hardwired designs (all of which 'should' have worked), I had no choice, and at last one design worked.
I spent ~3 h earlier today soldering two different circuits (one that I know will work -- using a dual op-amp (TLO82) with gain=100; and one that SHOULD work -- using a quad op-amp (LM324N) with gain in 3 stages for total gain= ~550, which is closer to the total gain I need); I'll get to test them on Monday. for my application, the secrets were (1) the first stage of the ap amp *HAD* to be a buffer (voltage follower), and (2) the excitation voltage of the load cell and the power supply for the op amp had to be from the same DC power supply (because of grounding issues?!?). (I had previously used a Boss PSA adapter, which provided 9.78 V DC, but I needed to dial in exactly 10 V DC for the excitation voltage.) thanks again for all of your collective help in this effort... at least I could reduce the error from using sub-par materials for this project! |
Edited by - pawnshop_trash on 11/06/2005 07:07:03 |
|
|
| |
Topic |
|
can anyone recommend LOW NOISE resistors, op amp?
