MkIIC+ inspired preamp.. DIYLC layout drafting

[Lowerleftcoast]

ultimately I decided it was redundant to have the 47nF coupling cap, followed by another 22nF coupling cap, especially with this 'lead channel always on' design where there isn't any switching.
so i dropped that 22nF cap completely. ... maybe it does matter? i don't really know.

Are you familiar with reducing a pot value by adding a resistor? For instance, a 1M pot with a 1M resistor soldered to terminals 1 & 3, will result in a pot with ~500k resistance. Essentially, it will function as a 500k pot.

By removing the 22nF cap, the previous 100k resistor is now acting as a parallel resistance with the 1M pot (of course there is the 680k and 470k resistors involved). Let's ignore the 470k for now and look at the 680k in series with the 1M pot. When you parallel the 100k resistance, the total resistance will be ~94k. When we figure in the 470k the total resistance can get down to around 90k.

So, removing the 22nF cap may not have much affect on the low frequency cutoff but the circuit will lose highs. IDK if this preamp would be better with less highs. Subjective.

yep, first dropping resistor, 10K, was dropped to *2.2K (just checked the layout)

and yes, the following B+ dropping resistors were changed to 3.3K (from 1K), as per a suggestion made my by a wiser poster, earlier on in this thread.
im not going to claim that i can explain why 3.3K is a better choice over 1K, but apparently it's for better filtering.

When AC is full wave rectified into DC there is a strong ripple current. It is double the 50Hz frequency from your power source (Wall voltage). 100Hz ripple will become *signal* and amplified by this preamp if measures are not taken to reduce the ripple. This will sound like a buzz or hum.

The 47uF reservoir cap smooths the ripple by charging and discharging 100 times per second (100Hz). When you get rid of the (10k or 2.2k) resistance after the reservoir cap the B+2 cap (22uF) is in parallel with the reservoir cap. These two caps in parallel make a capacitance of 69uF. This 69uF capacitance is now the reservoir cap. The potential problem that is created is that the B+2 components are now being powered with DC that contains the strong ripple current. There is no additional filtering at B+2 which would have been there if the resistance after the 47uF were still in the circuit.

The dropping resistors, drop voltage of course but they play another role. They form 1st Order Low Pass RC filters with the filter caps. If you like math you can calculate the desired frequencies to filter out the 100Hz content in the DC supply. I tend to just use an online low pass filter calculator to get the numbers.

A 1st Order Low Pass RC filter rolls off frequencies at 6dB per octave. The cutoff (or corner) frequency starts at 3dB below the cutoff frq. If we choose a cutoff frequency of one octave below 100Hz (=50Hz), the 100Hz content will be 9dB (6dB+3dB) quieter. (An octave lower is half the Hz.)

Using a low pass filter calculator with the B+2 original values. R=10k C=10uF the cutoff frequency is 1.7Hz. How many octaves is that below 100Hz?
100/2 = 50
50/2 = 25
25/2 = 12.5
12.5/2 = 6.25
6.25/2 = 3.12
3.125/2 = 1.6
It looks like about 6 octaves.
At 6dB per octave, plus that 3dB... it is close to a 39dB reduction of the 100Hz content at the B+2 node.

Just don't understand what he means by his channel version with only pin9 of V3.
But I can give it a try as described here.
Note that he has lowered the voltage a bit from 70V to 48V.

On this 12.6V heater supply, it appears he is using the impedance of the filament of V3 to replace the two 100R resistors. As you know the filament is connected from pin9 to pin4 and to pin5 on a 12A_7.

Elevating heaters. The 70V or 48V is not critical. This preamp has a cathode follower (V3b). The cathode voltage is comparatively high on this cathode compared to the other cathodes in this amp. The tube manufacturers have a rating stating they can handle around 100v difference (depending on the tube). To hopefully keep heater hum low, elevating the heaters so the difference is not near 100v is the goal.

 

[owlexifry]

Are you familiar with reducing a pot value by adding a resistor? For instance, a 1M pot with a 1M resistor soldered to terminals 1 & 3, will result in a pot with ~500k resistance. Essentially, it will function as a 500k pot.

By removing the 22nF cap, the previous 100k resistor is now acting as a parallel resistance with the 1M pot (of course there is the 680k and 470k resistors involved). Let's ignore the 470k for now and look at the 680k in series with the 1M pot. When you parallel the 100k resistance, the total resistance will be ~94k. When we figure in the 470k the total resistance can get down to around 90k.

So, removing the 22nF cap may not have much affect on the low frequency cutoff but the circuit will lose highs. IDK if this preamp would be better with less highs. Subjective.

holy crap. yes. good point.

yep, quite familiar (previously w/ an OD pedal build i needed a 25k pot and used a 50k pot + 50k resistor)

had a feeling something wasn’t right, as soon as @Ocirnelooc asked about it, looking back at the schematic had me reconsidering and pondering what could be affected by omission of that cap…

many thanks for pointing that out.

highs are important.
i’ll be inserting that 22nf cap back in.

When AC is full wave rectified into DC there is a strong ripple current. It is double the 50Hz frequency from your power source (Wall voltage). 100Hz ripple will become *signal* and amplified by this preamp if measures are not taken to reduce the ripple. This will sound like a buzz or hum.

The 47uF reservoir cap smooths the ripple by charging and discharging 100 times per second (100Hz). When you get rid of the (10k or 2.2k) resistance after the reservoir cap the B+2 cap (22uF) is in parallel with the reservoir cap. These two caps in parallel make a capacitance of 69uF. This 69uF capacitance is now the reservoir cap. The potential problem that is created is that the B+2 components are now being powered with DC that contains the strong ripple current. There is no additional filtering at B+2 which would have been there if the resistance after the 47uF were still in the circuit.

The dropping resistors, drop voltage of course but they play another role. They form 1st Order Low Pass RC filters with the filter caps. If you like math you can calculate the desired frequencies to filter out the 100Hz content in the DC supply. I tend to just use an online low pass filter calculator to get the numbers.

A 1st Order Low Pass RC filter rolls off frequencies at 6dB per octave. The cutoff (or corner) frequency starts at 3dB below the cutoff frq. If we choose a cutoff frequency of one octave below 100Hz (=50Hz), the 100Hz content will be 9dB (6dB+3dB) quieter. (An octave lower is half the Hz.)

Using a low pass filter calculator with the B+2 original values. R=10k C=10uF the cutoff frequency is 1.7Hz. How many octaves is that below 100Hz?
100/2 = 50
50/2 = 25
25/2 = 12.5
12.5/2 = 6.25
6.25/2 = 3.12
3.125/2 = 1.6
It looks like about 6 octaves.
At 6dB per octave, plus that 3dB... it is close to a 39dB reduction of the 100Hz content at the B+2 node.

thanks for writing out this excellent and thorough explanation.

 

[andrewRneumann]

On this 12.6V heater supply, it appears he is using the impedance of the filament of V3 to replace the two 100R resistors. As you know the filament is connected from pin9 to pin4 and to pin5 on a 12A_7.

Yes, I don’t believe you need the 100R resistor virtual center-tap if you use V3 pin 9 to connect to your heater elevation circuit. In fact, they might cause more trouble because your heater circuit would then have two ground references instead of one. I would recommend doing one or the other, but not both.

 

[Ocirnelooc]

That's good to hear. So I can actually use option 2 without any problems?!
Then I don't have to add the two extra 100R resistors.
Does this technically make any difference? I've never seen it like this before...

 

[Lowerleftcoast]

Does this technically make any difference? I've never seen it like this before.

I am not familiar with utilizing the pin9 option either. I am not use to 12.6v heaters. The option would not be available with 6.3v heaters. I checked a few 12A_7 tubes and they have a DC resistance of ~10R for each *leg*. That may seem small but the construction of the filament may prove sufficient for the task. It seems like an elegant way to create an artificial CT.
Let us know how it works out.

 

[andrewRneumann]

That's good to hear. So I can actually use option 2 without any problems?!
Then I don't have to add the two extra 100R resistors.
Does this technically make any difference? I've never seen it like this before...
View attachment 1029097

Yes option 2 should work. Only connect v3 pin 9. The other pin 9’s should be left with no connection.

There should be no discernible audible difference between the two options, especially if you elevate the voltage like you are.

Let us know how it works out.

ditto.

 

[Ocirnelooc]

Thanks everyone for the responses.
Then I'm going to try it first by using the CT of the filament of V3.
I will post my findings.
Is there anything else I need to take into account with shielded cable? I was thinking from input to V1 pin 7 and from volume knob to V1 pin 2.
If you have any comments or suggestions on my layout, please let me know.

 

[owlexifry]

So, removing the 22nF cap may not have much affect on the low frequency cutoff but the circuit will lose highs. IDK if this preamp would be better with less highs.

thanks again for your feedback. i have now inserted this 22nF cap as indicated in the schematic.

out of curiosity, i decided to record a 'before and after' riff, with all controls/variables untouched/left set exactly the same, to try and get some sort of an objective comparison.

here's the results:

before (no 22nF cap)


after (with 22nF cap)


(yes, ill admit there's a lot of gain here and i'm also clean boosting it with an 808 overdrive, so i guess that would narrow things down a little...)

signal chain:
- epiphone explorer
- DIY OD808
- preamp
- 5 band EQ (new discrete 9 transistor circuit @ +33V)
- FX return peavey windsor (4x EL34, 100W)
- speaker out into fryette power station reactive load
- line out into focusrite scarlett 2i2
- recorded on garageband with mesa 4x12 cab sim

im not sure i can hear much difference, but i think there might actually be a little more top end detail (>7-8K?) coming through. i'll take that as a win.

either way, im happy, sounds good

 

[owlexifry]

Is there anything else I need to take into account with shielded cable? I was thinking from input to V1 pin 7 and from volume knob to V1 pin 2.

- i would also recommend using shielded cable for the leads going to and from the drive pot.

- make sure you only ground one end of the shielding (not both ends)

good tips here:

http://www.valvewizard.co.uk/Grounding.pdf

 

[Ocirnelooc]

I have another question regarding the 1M input resistor. In your layout you have them directly on the input connector.
Now it is the case that for my implementation I want an input on the front, but also on the back. Same principle as on the Mesa Boogie Studio preamp.
Does it matter much for the circuit where this 1M is positioned? Because I can only choose 1 input.
Or is it better that it is closer to V1?
Suggestions?

 

[owlexifry]

I have another question regarding the 1M input resistor. In your layout you have them directly on the input connector.
Now it is the case that for my implementation I want an input on the front, but also on the back. Same principle as on the Mesa Boogie Studio preamp.
Does it matter much for the circuit where this 1M is positioned? Because I can only choose 1 input.
Or is it better that it is closer to V1?
Suggestions?

yep, the way i've done it (with 1M grid leak resistor directly on the input jack) won't work for your split input arrangement.

as shown in your schematic, the 1M can go anywhere after the point where the x2 input jack tips merge, and before the grid pin.

i dunno what's going on with your build, but if it was me, i'd have that 1M resistor closer to V1, possibly directly on the grid pin to a dedicated nearby grounding tab.
(not sure if that's best practice, if not, then ground the 1M to an input jack sleeve (like everything else with the star grounding arrangement ive used in my layout).
i guess it also depends on whats going on with your grounding scheme.

 

[Ocirnelooc]

Circuit boards are in production. So I'll have to wait a little longer before I can turn on the soldering iron.
I have one more question/comment regarding your applied resistors.
And I'm referring to the plate resistors. I see that you have taken larger wattage variants for this. Maybe 1W resistors?
But when I calculate back (with your voltage values) I arrive at the following powers for the plate resistors:
V1a - 192mW
V1b - 177mW
V2a - 161mW
V2b - 111mW
V3a - 196mW

Now I planned to apply 500mW resistors everywhere. Except for the 2W and 3W resistors.
But as I reasoned it is not necessary to apply 1W resistors. Did you study this too?

 

[owlexifry]

Circuit boards are in production. So I'll have to wait a little longer before I can turn on the soldering iron.
I have one more question/comment regarding your applied resistors.
And I'm referring to the plate resistors. I see that you have taken larger wattage variants for this. Maybe 1W resistors?
But when I calculate back (with your voltage values) I arrive at the following powers for the plate resistors:
V1a - 192mW
V1b - 177mW
V2a - 161mW
V2b - 111mW
V3a - 196mW

Now I planned to apply 500mW resistors everywhere. Except for the 2W and 3W resistors.
But as I reasoned it is not necessary to apply 1W resistors. Did you study this too?

it is widely agreed that using higher wattage resistors (particularly plate resistors) results in lower hiss/noise.

low noise is important to me, so i ordered and used 2W metal film plate resistors.

there would no issue with using 1/2W resistors if that is your preference (although you'd be better off with at least 1W).

 

[Ocirnelooc]

Okay, thank you.
Then I use 1W plate resistors. To be continued..

 

[Ocirnelooc]

Small update:
I received the PCBs today. I immediately started soldering.
So far this is the result. I still have to get some resistors and order the transformer. I am also working on a design for the 5 band equalizer because I want to integrate it immediately.
To be continued..

 

[Ocirnelooc]

Update of the front plate

 

[owlexifry]

Update of the front plate View attachment 1055094

holy crap, that's tidy.