Spectrum Analyzer Test - 6PS vs 715P Capacitors

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

Some people believe that different capacitor types have a different type of sound and some don't. I have seen many tests online proving that there isn't any difference, however people still believe there is a difference. There is only one flaw I can find in these tests and that is the capacitors are not being subject to high voltages.

So as I've not been happy with the tone of my TW Express I thought I'd replace all the 715p Polypropylene Capacitors with 6PS Polyester Capacitors which are the correct type to use. I was also not happy with the silver mica capacitors I used as they were 470pF instead of 500pF (picky I know) and they were an unknown brand. I took readings before and after the capacitor replacement with a spectrum analyzer and obviously used my ears.

I will say now that I didn't actually replace the two coupling caps in the output stage as I plan on replacing the PI Snubber Cap when it's replacement arrives so thought I'd do all of that at the same time. Also I haven't replaced the 470pF Bright Cap as I plan on rewiring the amp with thicker wire and want to redo the bright switch at the same time so the caps are glued to the switch, all at a later date.

The 6PS Caps values were very slightly different, the 715p caps were .1uF, .022uF and .0022uF, whereas the 6PS Caps were .1uF, .02uF and .002uF and obviously the silver micas were changed from 470pF unknown to 500pF Cornell Dubilier. .02uF is within the 5% tolerance of the .022uF Caps but unfortunately as I don't have a capacitance meter I couldn't be precise and determine what they actually measured at.

Anyway I've attached the results and it turns out, there is a difference! In the 1.5Khz test the odd order harmonics have increased in volume and in the 3Khz and 6.1Khz tests a lot of the higher frequency harmonics have increased in volume. The latter result I find odd as the general consensus is that Polyester Capacitors roll off the high end a bit but in this case it has increased. As for the listening test well the amp sounds a lot better, the high end is no longer piercing and you can definitely hear an increase in the harmonics. The tone is fatter and much more pleasing to the ear!

So, what do you guys think?
You do not have the required permissions to view the files attached to this post.
User avatar
roberto
Posts: 1841
Joined: Tue Nov 27, 2007 4:45 pm
Location: Italy

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by roberto »

Nice stuff! I will ink this thread in other forums too.
Thanks!
User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

I should add that all controls were set at 12 o'clock (5 on the amp) and the bright switch was in the brightest position (currently 470pF). Glad you liked it!
User avatar
Aurora
Posts: 762
Joined: Mon May 26, 2008 7:51 am
Location: Norway - north of the moral circle!

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Aurora »

Not sure if I understand this correct, but how many caps are changed between the two test? Tolerance may play a role in the measurements. What are the tolerances of the two types you used? One example - the change from .022 to .02 is 10% in itself, - tolerance may skew that even more.
One thing that would make things clearer would be stacked plots in different colours, but I don't know which version of Pico you have...?
User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

Sorry should have been a bit clearer on that. The Tolerance for both sets of cap is 10% and 6 caps were replaced in total. Yes values may have been different but in theory that would only affect the lower frequencies? Also the difference between .02uF and .022uF is so small that you wouldn't normally notice. Then again I'm not sure how the tone control capacitors would have affected the harmonics. When I change the rest of the caps I'll take another set of readings.
User avatar
Merlinb
Posts: 233
Joined: Mon Nov 17, 2008 9:52 pm
Location: Burnley, Lancashire
Contact:

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Merlinb »

So you replaced six caps with ones of wildly different values and materials? I'm not surprised you saw a difference- the frequency response will have changed a bit. Probably nothing do with dielectric materials.
Also, what do you mean when you say the test is "100Hz 1.5khz" or whatever ? Did you apply two tones? What amplitudes? They're clearly not the same...

While polyester dielectrics do produce more distortion than other plastics, you will ever detect in in a guitar amp, for the simple reason that the distortion of the valves is FAR FAR FAR higher!
vibratoking
Posts: 2640
Joined: Tue Nov 10, 2009 9:55 pm
Location: Colorado Springs, CO

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by vibratoking »

Anyway I've attached the results and it turns out, there is a difference! In the 1.5Khz test the odd order harmonics have increased in volume and in the 3Khz and 6.1Khz tests a lot of the higher frequency harmonics have increased in volume.
This is interesting. Thanks for sharing. It would help to understand some more details of how you made the measurement. SA attached to the speaker output? If so, what is the load? A speaker or something else? What does the input spectrum look like? Network analyzer might be a better tool, but use what you have. Why did you choose 250mV as the input amplitude? It looks like the input is 100Hz in all cases? If it were me, I'd be testing at 400Hz instead. More in the typical range of the guitar. What do you mean by a 1.5kHz test vs a 3kHz vs a 6.1kHz test? Is this the sweep? If so, are you adjusting the RBW for each test? Why are you using different sweeps? Not sure changing the sweep is useful or necessary. Lot's of questions to be asked, but it's hard to pin them down without knowing the test details.

Overlayed plots would be really helpful as well as consistency in the plot scales.
Electronic equipment is designed using facts and mathematics, not opinion and dogma.
User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

Sorry again should have been clearer.

The signal used was 100Hz, 1.5Khz is the max frequency I measured on that test. So the other tests were 3Khz and 6.1Khz. Surely the slightly different values wouldn't affect the high frequencies though? Especially in the Khz range?

250mV was the lowest I could go unfortunately on my Picoscope. I did tests measuring up to 1.5Khz, 3Khz and 6.1Khz as its easier to see the lower frequencies on the 1.5Khz range and with the 3Khz I thought I may as well as I had the option.

This is the first time I did a test like this, I didn't mean to piss anyone off.

Edit: Also the caps aren't 'wildly' different values, its not like I replaced a .1uf with a .022uf cap. All .1uf caps were replaced with .1uf, .022uf replaced with .02uf and .0022uf replaced with .002uf.
Last edited by Littlewyan on Mon Dec 01, 2014 8:29 pm, edited 1 time in total.
User avatar
Merlinb
Posts: 233
Joined: Mon Nov 17, 2008 9:52 pm
Location: Burnley, Lancashire
Contact:

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Merlinb »

Littlewyan wrote:Surely the slightly different values wouldn't affect the high frequencies though? Especially in the Khz range?
I don't know what caps you changed, but if they alter the amplitude of the 100Hz fundamental as it passes through the amp, then you should expect the harmonics to be altered too, since the valves are handling a slightly different signal amplitude.
Also, the caps in the amp influence the various dynamic changes in valve bias during use. Changes in capacitor value will change these bias shifts, and so change the distortion character of the valves.
vibratoking
Posts: 2640
Joined: Tue Nov 10, 2009 9:55 pm
Location: Colorado Springs, CO

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by vibratoking »

Littlewyan wrote:Sorry again should have been clearer.

The signal used was 100Hz, 1.5Khz is the max frequency I measured on that test. So the other tests were 3Khz and 6.1Khz. Surely the slightly different values wouldn't affect the high frequencies though? Especially in the Khz range?

250mV was the lowest I could go unfortunately on my Picoscope. I did tests measuring up to 1.5Khz, 3Khz and 6.1Khz as its easier to see the lower frequencies on the 1.5Khz range and with the 3Khz I thought I may as well as I had the option.

This is the first time I did a test like this, I didn't mean to piss anyone off.
I don't believe you are pissing anyone off. On the contrary, I believe you are generating interesting information. At least it's interesting to me. Unfortunately, this falls into the "no good deed goes unpunished category'. I don't think you can avoid this situation. With measurements like this, the devil is in the details. If you don't sort the details, you are just making measurements that could lead to correct or incorrect conclusions. You will never know the difference between the two without diving into the details.

If it were me I would always be sweeping the full audio range or even twice the audio range. I would zoom on the data afterwards. I assume you can output the data digitally? Make sure the RBW is set properly so that you are not 'smearing' frequencies. It seems the RBW is fine enough. How many averages are in your plots?
Last edited by vibratoking on Mon Dec 01, 2014 9:00 pm, edited 1 time in total.
Electronic equipment is designed using facts and mathematics, not opinion and dogma.
User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

I guess I took what you guys were saying the wrong way, apologies. Attached is a screenshot of the schematic with the caps that I changed circled.

The Spectrum Bins was set to 1024 which according to Pico Technology gives me a refresh rate of 204.8ms and a frequency resolution of 4.883Hz. Sorry I don't know what you mean by averages. The analyzer I used is also an oscilloscope, it plugs into my computer using USB and I use software to control it.
You do not have the required permissions to view the files attached to this post.
User avatar
gktamps
Posts: 741
Joined: Sat Nov 20, 2010 3:05 pm

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by gktamps »

Interesting thread!
You aren't pissing anyone off. You have simply attempted to to shift away from opinion to science, and in doing so and presenting data, your methods will be examined with great scrutiny, as is scientifically appropriate.
Study design, instrument selection, operational methods, and a priori assumptions all accounted for and being appropriate make the test results relevant and repeatable.
Please persist.
matt h
Posts: 1224
Joined: Fri May 12, 2006 2:26 am
Location: New England

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by matt h »

(deleted)
Last edited by matt h on Fri Mar 27, 2015 7:10 am, edited 1 time in total.
User avatar
Phil_S
Posts: 5935
Joined: Tue Oct 23, 2007 10:12 pm
Location: Baltimore, MD

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Phil_S »

I think we simply have some hard-nosed probing going on to
a) understand what was done and
b) to ascertain if there was adequate scientific rigor.
This is a good thread. Depending on what you did, you may have to rewind the test(s) at some point to improve the data.

I don't know enough to do this, but I've been telling myself for years that someone should be able to prove the claims being made about caps.
User avatar
Littlewyan
Posts: 1911
Joined: Thu Sep 12, 2013 6:50 pm
Location: UK

Re: Spectrum Analyzer Test - 6PS vs 715P Capacitors

Post by Littlewyan »

Well my intention was to answer this question once and for all. Do capacitor types make a difference. However as I didn't measure the values exactly its still not a great test but I personally believe they do.
Post Reply