Debugging the Sparkle Boost

[R.G.]

I have moved this here from the "What to do when it doesn't work"
sticky to avoid muddying up the waters over there.

===========================================
Posted: Thu Feb 03, 2005 4:50 pm    Post subject:    

--------------------------------------------------------------------------------

R.G. wrote:
OK. We have a sample case.
=============================================
Quote:
1.When I turn it on, the volume is cut in half. The gain knob is unresponsive, and the volume pot does work.
2.Dragonfly's Sparkleboost
3.http://www.diystompboxes.com/sboxforum/viewtopic.php?t=28818 ; http://img.villagephotos.com/p/2004-7/770357/SPARKLEBOOSTPERF.gif
4.No mods  

OK so far. But
Quote:
5. 3K9 subbed for 2K7, 10K pot for 5K,2.4M for 2.2M, MPF102 for NTE458, no indication LED  

As I said, the first question I'll ask is if you checked the pinout of the subtituted part against the original.

I did. See http://pubpages.unh.edu/~aperkins/pdf/MPF-devices/MPF102.pdf for the datasheet, and notice that whether or not the MPF102 and the NTE458 are the same pinout, the MPF102 is pinned GSD and the layout is for SGD. If it is wired exactly like the layout, the gate and drain are reversed. A quick look at the NTE458 has the SGD pinout.

Did you allow for that when you wired it up?

Quote:
6.No
7.No battery  

? You need a battery. Does this mean that you are powering it from a DC adapter in spite of the admonition to use a battery?
Also, you have not supplied the voltages of the red and black wires from (what should be) the battery. These are important, as they verify that the battery voltage has made it onto the board OK. We really need that info.

Quote:
Q1;MPF102
D = 4.78 (its as close I could get it, it should be 4.5v)
G = 0.06
S = 2.02  

Hmmm... I'm trying to think what would happen in that circuit if the gate and source were reversed. It ... might... give you those voltages.

Quote:
Meter set to 20k ohm; no power to circuit for these measurements
D = 4.24
G = 0.76
S = 4.09

I'm not sure why you did these. I only asked for voltages. Good commentary though - I'l clarify that.

So - can you check the things where I asked for more?


Here's the info I forgot to put in there.


DC jack voltages(yes, I am using a wall wart....  ) If you really think it'll make a difference, I'll wire up a battery clip and put in a fresh battery later.

positive wire to board- 9.53v
negative lug(s)-0.00v

I swapped the S and G so I wouldn't have to modify the circuit for the MPF102, so yeah, I checked and confirmed the pinouts.

I'm just not quite sure why I don't have a voltage reading for the Gate leg. I don't know if it is good or bad. The power supply seems to be fine. It should be, I bought it brand new about a week or two ago.
_________________
~Rj

OK, now we have a set of info we can work with. The following is a reasonably exemplary set of reasoning that we would do on any such setup.

positive wire to board- 9.53v
negative lug(s)-0.00v

Check! Good voltage to the board.

Q1;MPF102
D = 4.78 (its as close I could get it, it should be 4.5v)
G = 0.06
S = 2.02  

We have an N-channel JFET (datasheet on the MPF102 verifies this) and the pinout has been checked and is correct per the linked drawing. So now we compare the voltages against what must be there for an N-FET to work.
(By the way - for observers of voltage-based debugging, what follows is standard for JFETs, and is not a particularly good example for opamps and bipolars. JFETs being so wide-parametered, you have to guess, and get answers as ranges, not dead-on numbers. -  R.G.)

Looking at the MPF102 datasheet, I see that Vgs for the MPF102 is specified at -0.5 to -7.5V for an Id of 200uA. Oh, goodie. A Vgs hole you can drive a truck through. Well, that accounts for why you need a 100K pot to tune it in, I guess. For an N-channel JFET to be working as a linear amplifier, there's a different set of data than for a bipolar. The source must be more positive than the gate, and the amount of that more positive has to be less than Vgsoff. The drain of the JFET has to be lower than the power supply and have some voltage between it and the source to wiggle around. If it's at the source voltage, the JFET is saturated, and not much if any signal gets through.

We have: source at +2V, gate at 0V - that's OK for an N-channel. The drain is at +4.7V, again OK.

The bias voltages are at least plausible for an N-channel JFET, so the device should be amplifying (if I haven't missed anything) and that says to look at some other parts.

Possibilities are that there's a wrong part value, a wrong polarity, or a manufacutring problem. So we go look there.

RJ - I need you to do the following:
Measure the resistance from gate to ground, but use the red meter lead on ground, not the black one for this measurement only.
Measure the DC voltage at both ends of all three capacitors, and the terminals of the pots.

Since the device seems to be biased right, we look for what could be lowering its gain dramatically.

[puretube]

low Z loading?

[RjM]

Resistance test, Gate to Ground- 0.82            Meter set to 20K Ohm

1uF cap- 5.03

10uF cap- 2.04

.022uF cap- 0.10

[R.G.]

low Z loading?

Yeah, pretty much gotta be that, or a high resistance out. Just gotta poke around and find it.

Resistance test, Gate to Ground-0.82 20K Ohm

I realize I did not tell you to turn off the voltage for that measurement.

If you did, great. Change it to a 1M or greater scale and try it again. What you have here says that the gate is being pulled dow with a low resistance, which could be the problem, but we need to be sure that we found what we think.

1uF cap- 5.03
10uF cap- 2.04
.022uF cap- 0.10

Sorry - again I wasn't clear enough. I need this taken with power on, both sides of the capcitor. If one of them is 0V,  I need to know that - in fact, that's what I was looking for. One side should be at 0V, and if it's not, it indicates a problem.[/quote]

[RjM]

10uF (the only polarized cap)
pos leg - 2.01
neg leg - 0.00

1uF
leg 1 -0.03
leg 2 -5.14

.022uF
leg 1 -0.01
leg 2 -4.05 (It was weird with this one though, the number kept on going down until it was about 1.03-0.97) Does that mean it is a leaky cap or something? Maybe the imput cap is thinning, or weakening the imput signal, giving the transistor less to work with, like turning down the volume knob on a guitar...?

For the cap tests, I put the black lead to ground, and touched each leg of each cap. oops, I forgot about the pots. I'll go do that...


Gate to ground- .001    Meter set to 2M ohm

Yes, I didn't have power to the circuit during the resistance test.

[R.G.]

OK. We have something to go on now.

You said:

Gate to ground- .001 Meter set to 2M ohm
-and-
Gate to Ground-0.82 20K Ohm

What that tells us is that the gate is being pulled to ground by a resistance that measures something like 820 ohms. That would be definitive as a problem if it weren't for the possibility that the meter is measuring the forward conductance of the gate/channel diode.

Measure the gate to ground with the red/black meter leads one way and then reversed, on the 2M scale. If one way or the other is not 1M or more, we have found at least one of your problems.

10uF (the only polarized cap)
pos leg - 2.01
neg leg - 0.00

1uF
leg 1 -0.03
leg 2 -5.14

Those all look good. Paticularly the voltage being higher on the + end of the electro than on the - end.  

.022uF
leg 1 -0.01
leg 2 -4.05

There's a problem. The 0.22 has a resistor to ground on both sides. It should read 0V on both leads.

With reference to the posted layout, which lead is lead 1? Left or right?

By the way, start looking for why the gate is showing too low a resistance to ground. Is that "1M" resistor really a 1M? Is there a solder short under the board, stuff like that.

[RjM]

I forgot to tell, but I subbed a 820k for the 1M.... wait.... HAHAHAHAHA!!!! I accidentally subbed a 820 ohm for the 1M... definately way too low! Oh man, you called it.

Here are the resistance measurements.

Red to gate,Black to ground- .000

black to gate, red to ground- .001

leg 1 is the right, and leg 2 is the left, in reference to the layout.

[R.G.]

I forgot to tell, but I subbed a 820k for the 1M.... wait.... HAHAHAHAHA!!!! I accidentally subbed a 820 ohm for the 1M... definately way too low!

Ding!!

That goes a long way toward explaining the low volume.

.022uF
leg 1 -0.01
leg 2 -4.05

leg 1 is the right, and leg 2 is the left, in reference to the layout.

Hmm... that's odd. How's come the **outside* of the 0.022uF is up at 4V???

[RjM]

Hmm... I think I just noticed another problem. Alright, look on the layout. Where the wire that connects to the middle lug of the 5K gain pot, I have the wire over two dots. Would that change anything? Probably one of the reasons why the gain knob doesn't work... I'm going to go ahead and assume that I should have it before the wire. (the one connecting to the very right lug of the 5k pot)

I can't get a 1M ohm resistor now, could I use a 2.4M ohm resistor, or would that be too much?

[R.G.]

I can't get a 1M ohm resistor now, could I use a 2.4M ohm resistor, or would that be too much?

2.4M will be fine.

I think I just noticed another problem. Alright, look on the layout. Where the wire that connects to the middle lug of the 5K gain pot, I have the wire over two dots. Would that change anything?

As long as it connects to any of the dots with that blue line along the bottom, it's fine.

[RjM]

It looks like I found yet another mistake, but through careful inspection, it seems to be the final problem. I just need to fix it as soon as I get home, and I'll have a working sparkleboost! It'll be great. By the way, what is a feedback loop when using diodes? I know it has something to do with distortion.

[R.G.]

By the way, what is a feedback loop when using diodes? I know it has something to do with distortion.

?

You can make distortion by using diodes any of a zillion ways in or around a feedback loop.

Feedback is just that. You have some circuit that amplifies. You connect some parts between the output and the input to feed a metered amount back to the input. Feedback may be negative, where it lowers the gain you'd otherwise get, but also stablizes things and reduces the sensitivity of the circuit to exactly what the inside amplifier does; or positive, where it increases the gain of the circuit, and also  increases the sensitivity to erratic stuff in the innards. Both have their uses, but we would not want to live in an electronic world without negative feedback to stabilize things.

[RjM]

Well, the gain pot works now, but it just seems to quadruple the volume. There still seems to be something wrong. I got some soundclips that I'll upload here in a second. Also, if part of the cover of a resistor was scratched off, and you could see the metal casing, and part of the resistor material, would that make that resistor bad? Could it explain the weird voltage on the left lead side of the input cap?

Well, I don't know where I could upload soundclips, so I'll explain what it is doing. When I turn the volume pot to 0, and then I start to turn it up, there is this massive humming/buzzing where no guitar is present, and then when I turn it up after that, the humming/buzzing is gone, and you can hear guitar.

When I do the same with the gain knob, near almost having it up all of the way, it quadruples the volume. The volume still isn't at normal levels though. Also, when resistors are connected in series, do their values add up? I have two 47k resistors in series subbed for the 100k resistor. Dang it, that's another sub I forgot to mention, but there aren't anymore part substitutions that I forgot to mention. I tried plugging the sparkleboost in chain with my MT-2, and when I did that, there was no sound. The output cord from the sparkle went into the input of the MT-2.

The transistor voltages are all weird too...

D- 4.87
G- 4.53
S- 4.71

[NaBo]

oo oo, i know this one...  check your grounds  

[RjM]

What should I be looking for? A disconnected wire? How can I use my DMM to pinpoint the problem? I used my DMM to check for continuity, and the resistance was very low, suggesting that you're right about the ground problem.

[petemoore]

Ground is a reference voltage for the circuit. All points ground, with battery connected should read 0v potential from ground.
 All points ground should read 0 resistance from ... or ... continuity to ground.

[petemoore]

On the first post there is a schemtic link to the Sparkle Boost...at the top left is a 'part'...is that a PS Filter cap?
 Also I read about 'diodes' in a Sparkle Boost thread...what is that...diodes to ground added from signal path near the output ?

[Dragonfly]

On the first post there is a schemtic link to the Sparkle Boost...at the top left is a 'part'...is that a PS Filter cap?

2.1mm dc jack, back side....

[RjM]

Ground is a reference voltage for the circuit. All points ground, with battery connected should read 0v potential from ground.
 All points ground should read 0 resistance from ... or ... continuity to ground.

The reading from the DMM is near enough to 0 resistance to assume that the grounds are all fine, and I don't think that the effect would work at all if a ground was disconnected. What is your take on the crazy transistor voltages?

D- 4.87
G- 4.53
S- 4.71

[petemoore]

Ground is a reference voltage for the circuit. All points ground, with battery connected should read 0v potential from ground.
 All points ground should read 0 resistance from ... or ... continuity to ground.

The reading from the DMM is near enough to 0 resistance to assume that the grounds are all fine, and I don't think that the effect would work at all if a ground was disconnected. What is your take on the crazy transistor voltages? >>>Do you care to share the DMM reading
 >>>What makes you assume that ? Connect ground to ground, and 0v potential between the two ground points should be found.
 Look for continuity, or alternatively 0ohms resistance between ground and GND. Try test clips on a wire to 'get' ground when in doubt.
D- 4.87
G- 4.53
S- 4.71

The transistor isn't within operational range.