Where did my bass go?

[ficklecycler]

Hi there,

I just got finished building my GSG. Overall, it's quite nice. It's got great mids and highs, but I am unsatisfied with the bass. I'm noticing that most people here aren't experiencing this problem so I'm thinking it's an issue with my build. Let me be clear, I know there is a limit to bass production with inexpensive iron and all that, I'm not expecting REALLY high output. That being said, compared to my other tube amp, the GSG is sounding really unbalanced in the bass department.

Someone on another great audio forum (karma) has been helping me with the problem and offering some insight as to why my GSG might be putting out less bass. This person suggested that I might try a 3.5K transformer vs the 5k. It's been nice having a little feedback to go off of so if anyone here can help me out, I'm all ears.

I'm using some NOS 6B4G's from Holland, Edcor GXSE15-8-5K outputs, and a XPWR005-120 Output. The rest of the build is by the book from the supply list. My speakers are Audio Nirvana Cast 8's. I find the bass on them to be satisfactory with most other amps.

Any thoughts?

[Shannon Parks]

Welcome, and congrats on your get*set*go. We'll get this figured out.

First, have you compared your voltages to the voltage chart in the manual? That would be helpful if you can post that.

Second, could this be a phase cancellation issue? Just swap the '+' and '-' on one of your speakers to test for it.

Shannon

[ficklecycler]

Hi there, thanks for helping out. I'll get the measurements on Saturday when I have some free time and post them up here.

[ficklecycler]

Hi guys, I had some time to go through the amp and make some measurements. I shorted the inputs, and I left test speakers connected.

I made a nicer table in GMAIL and all of the pin numbers didn't transfer over. Is this still readable to you? The readings are all in the exact same order as the manual. I'm noticing that my smaller AC voltages are WAY WAY off. I made sure I had my multimeter set on the right setting but maybe something is messed up? Again, thank you for your help.

V1

0
152.1
1.9
0
162.5
1.8
108.4 50.1 Bias
108.4 50.1 Bias

V2

0
111.3 51.6 Bias
338.5
0
338.5
0
111.1 51.6 Bias
344

V3

0
47.3
324
0
0
0
53.6
0

V4

0
52.1
323
0
0
0
46.1
0

J1 & J2 0.0 all around

J3

112.1 51.3 Bias
112.3 51.3 Bias
0
53.3
47.1

J4

109.5 50.0 Bias
109.3 50.0 Bias
0
51.9
46.1

J5

346
340

J6

341.5
0
0
340.5

J7

330
330

[Ty_Bower]

Your meter is just getting confused trying to measure AC with the DC bias on top of the signal. All your voltages look just fine except for one. Can you double check the voltage at pin 3 of V4? That is the plate voltage for the output tube, and you are showing only 52.3 volts DC. That should be over 300 volts DC.

[Shannon Parks]

And a couple of nice big interior pix can help, too, which really helps us spot wiring errors. If you can't up them to Flickr, by all means email them directly to me and I'll post them.

Shannon

[ficklecycler]

Your meter is just getting confused trying to measure AC with the DC bias on top of the signal. All your voltages look just fine except for one. Can you double check the voltage at pin 3 of V4? That is the plate voltage for the output tube, and you are showing only 52.3 volts DC. That should be over 300 volts DC.

I copied it down incorrectly, pin 3 has 323 Volts on it. Do most meters get confused, or is it just because mine's a cheapy?

[ficklecycler]

And a couple of nice big interior pix can help, too, which really helps us spot wiring errors. If you can't up them to Flickr, by all means email them directly to me and I'll post them.

Shannon

Sure thing. This shot is before the last few wires were connected (this pic is just to show the wiring up top.)


I'm going to ground the chassis once I understand exactly how. Nothing has really changed since this next picture.


...oh and yes those are earplugs shoved under the caps. I kept bumping them and bending the leads so those gave them a little push up support haha. I took them out after I was done in there.

[Shannon Parks]

Thanks, ficklecycler for the voltages and pic. The voltages look good.

1) Did you do a phase cancellation test? Just listen to one channel at a time, or swap one of the speaker connections.
2) Check your RCA input connectors. Use your DMM to be sure that the center posts go to J1-1 and J2-1 respectively.
3) Make sure R9 and R10 are 475K. You can measure them with the amp off with your DMM. Also double check that C1 and C2 are .22uF (though they look it).

Shannon

[EWBrown]

What value and voltage ratings are C5 and C6 on your GSG board? They should be 100 uF / 100VDC.

If the capacitance values of these two (C5 and C6) is too low, then the bass will be weak, and distortion will increase significantly with increasing volume levels and power output, as the tubes will fall out of clean Class A operation, and slip into a very "messy" Class B condition, which is not a good thing

The DC voltage measurements acros the 6B4Gs' RKs R15a, b, c and R16a, b, c (or on TP1 and TP2) will read correctly, in an idle (or no signal input) condition, even if C5 and C6 were under-valued or even if they were missing.

Also check the resistance from TP1 to ground, and TP2 to ground (use J6, point 2 or 3 for this), when the amp is powered off, and all of the power supply caps are fully discharged , it should read very close to 900 ohms (3 2.7Ks in parallel, for each channel), from the TPs to ground.

With approximately 55 to 60 mA plate current for each 6B4G, this would then properly read between 50 and 55VDC from TP1, or TP2, to ground.

Using these two Test Points will read the average DC voltage between pins 2 and 7, in your case, 48 and 54VDC, which would then be 51VDC, which is very good, at around 58 mA plate current, read across 900 ohms, for each of the 6B4Gs.

Also, if you have not done so already, the unused ground point on terminal strip J6 (2 or 3) should be connected to a physical chassis ground.

One of the power transformer mounting screws would be a good chassis grounding point.

Also, the third wire "ground" on the AC power input connector should also be connected to the chassis at this location This would be for electrical safety reasons.

/ed B

[ficklecycler]

Thanks, ficklecycler for the voltages and pic. The voltages look good.

1) Did you do a phase cancellation test? Just listen to one channel at a time, or swap one of the speaker connections.
2) Check your RCA input connectors. Use your DMM to be sure that the center posts go to J1-1 and J2-1 respectively.
3) Make sure R9 and R10 are 475K. You can measure them with the amp off with your DMM. Also double check that C1 and C2 are .22uF (though they look it).

Shannon

I did do a phase cancellation test. Everything seems fine there.

Used the DMM and double checked the RCA's. No problems there either.

R9 and R10 are the correct values, as are C1 and C2

Just a side question. How would switching to a 3 or 3.5k transformer change the sound signature compared to the current 5k?

[ficklecycler]

EWBrown, my bold type isn't yelling at you FYI

What value and voltage ratings are C5 and C6 on your GSG board? They should be 100 uF / 100VDC.

Double checked and they are the correct values


Also check the resistance from TP1 to ground, and TP2 to ground (use J6, point 2 or 3 for this), when the amp is powered off, and all of the power supply caps are fully discharged , it should read very close to 900 ohms (3 2.7Ks in parallel, for each channel), from the TPs to ground.

Both sides measured 901 ohms


Using these two Test Points will read the average DC voltage between pins 2 and 7, in your case, 48 and 54VDC, which would then be 51VDC, which is very good, at around 58 mA plate current, read across 900 ohms, for each of the 6B4Gs.

Also, if you have not done so already, the unused ground point on terminal strip J6 (2 or 3) should be connected to a physical chassis ground.

Also, the third wire "ground" on the AC power input connector should also be connected to the chassis at this location This would be for electrical safety reasons.

I will do both of these things. Thank you for your help.

[Shannon Parks]

I did do a phase cancellation test. Everything seems fine there.

Used the DMM and double checked the RCA's. No problems there either.

R9 and R10 are the correct values, as are C1 and C2

Just a side question. How would switching to a 3 or 3.5k transformer change the sound signature compared to the current 5k?

Thanks for confirming those items.

I don't think your problem is the 5K output. I did simulate a get*set*go yesterday with a 5K output and the power output dropped a little bit while distortion dropped a little bit, too. I was getting sub-20Hz with a triode strapped 6L6 with that exact transformer, so a 6B4G - which its smaller plate resistance - will be sub-20Hz easily.

Shannon

[Shannon Parks]

I think at this juncture I would want to look at the tubes and also see if you could test it.

1) You're from San Francisco, so *you must* have a smartphone, right? Get a signal/function generator app and we'll test this baby out. Be sure the app can do 20Hz. You want to test with a 20Hz and 1kHz sine wave at 200mV. Your DMM may not be True RMS - i.e. able to read AC voltages outside 400Hz to 2kHz accurately. The phone app may also be inaccurate at low frequencies due to the output coupling capacitor used. But those things don't matter. We will make relative measurements. Disconnect your speakers. Preferably test with 8 ohm loads, but you can do these tests unloaded since you are testing SE outputs. The gain would then appear a little higher, but once again, we are looking at relative measurements.

Short right channel and test the left.
a) 1kHz, 200mV - Measure voltage at J2-1 (should read around 200mV AC).
b) Measure output voltage. It should read around 1.5 to 2V. Divide that number by the *measured* input voltage to see your gain.
c) 20Hz, 200mV - Measure voltage at J2-1 (should read around 200mV AC).
d) Measure output voltage. Divide that number by the *measured* input voltage to see your gain.
e) Do the other channel, looking at J1-1.
f) Post those numbers here.

You can figure out dB loss: 20log(gain20Hz/gain1khZ).

Email me directly with any specific questions regarding this test, too.

2) Can you swap your tubes with others? If not, I can send you some to try out.

Your build looks spot on. Maybe email me a original resolution version of that final pic, if you can.

Shannon

[EWBrown]

EWBrown, my bold type isn't yelling at you FYI

I use the same "technique" sometimes, or use color or italics, as well. see below


I built two G*S*Gs with 5K Magnequest RH60 OPTs, and they both sound very good. The only slight trade-off is that the maximum output power is a bit lower, about 2.5WPC, but then the 2HD and 3HD are significantly lower, which is a very good "trade-off", IMHO

The bass is as good as ever with these trannies.

The RH60s are MQ's "entry level" SE OPTs, and are very nice for the price. These have 60 mA (continuous) rated primariwes, and I had inquired of Mike if these would eb sufficient for 60 mA constant duty, and he replied that these are good for that, and that they are very conservatively rated.

The audible difference between 2.5W and 3 to 3.5W max output is just barely discernable to most folks.

http://www.commonsenseaudio.com/nirvana.html

I haven't used or heard the Audio Nirvana Cast 8s, but if memory serves me correctly, they are supposed to be fairly similar in regards to SPL efficiency and freq response, as the 8 inch Fostex FE206Es, which are pretty decent performers in properly matched BLH enclosures, as are their smaller 6.5 inch relatives, FE166Es which I use in Madisound BLH 16 enclosures

Of course, some folks on another website have come up with a magical mystical sonic enhancement for the 6.5 inch FE166E speakers (see also the "Advanced Snake Oil" topic)

http://www.decware.com/newsite/DFR6.htm

Somehow, I just don't think that this technique would do ANYTHING at all to enhance the Cast 8s, in any way, form or fashion

/ed B in NC