18W TRex build - dark sounding

18watt-specific Tech Talk - Building, Fixing, Parts, Mods...

Moderators: CurtissRobin, colossal, zaphod_phil, Daviedawg, Graydon

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

18W TRex build - dark sounding

Post by SiRaff »

Hey everyone,

I've just put together a TRex as per Mark Huss' schematic, found here:
https://mhuss.com/18watt/schematics/18wTRex.pdf

I'm having issues with the Normal channel sounding very dark, even with tone control all the way up. We're not talking a little bit dark, but "blanket over the speaker" dark. Another way to see it is it's like the tone pot on the guitar is rolled back to around 5, which to my ears I would never use. I've tried with a Strat and an LP. The Normal channel is also not very gainy, only reaching crunch at around 8 on the volume. It sounds like there is some noticeable treble roll off, but I can't work out where it could be coming from. I've heard these amps are normally too bright if anything!

The Wreck channel sounds great to me (a fair amount of bass coming through which makes it mushy at higher volume, but I can tweak that with some component changes). It actually sounds fantastic at lower gain, especially with a Strat neck pickup. It has significantly more treble available than the Normal channel (which is unusable as things stand). In case this is relevant, I am getting a kind of "whoosh" sound on held notes when I have the volume up on the Wreck channel only - I don't mind the effect but haven't heard this before. I don't know if this is part of the amp's character (this is the only EL84 cathode biased amp I have) or if this indicates a problem somewhere.

It's worth saying that I'm using stock values on the components, and transformers from AmpMaker. The PT is non-centre tapped (so I'm using a bridge rectifier) but I don't see why this would matter.

Here's what I have tried, to no avail so far:
- Bypassing the Volume and Tone pots (and related components). I did this by desoldering the ground connections on both pots, then putting a jumper over from the plate of V1a to the coupling cap before the PI. No difference. (This should sound as though both controls are all the way up, right?)
- Removing the optional high pass filter around the tone stack (asterisked on the schematic). This shouldn't really affect the tone, since the knee is at 68Hz? I wanted to see if gain increased, but it's not that noticeable. Still lower than it should be.
- Changing speakers. I have two cabs available, a cheap Marshall speaker (from an MG50DFX, which is what I've been converting to make this build) - which is a bad speaker and may have some treble roll off. And an Egnater Rebel 112x cab. It COULD be that both speakers are bass heavy, but it's quite an extreme effect to be caused by speakers.
- Output Transformer. I'm using the 18W Push Pull offering from AmpMaker. I haven't swapped it, but I have another one of these in another build (the AX84 Lead preamp into 6V6 poweramp), and this amp has treble on tap when I need it. Admittedly the NFB with Presence control helps with that. I don't think it's the OT but it's possible.
- Lowering the coupling cap between V1 and the PI down to 5n (was 22n). I couldn't hear any difference.
- Removing the cathode bypass cap on the power tubes. Again, I couldn't hear any difference.
- Swapping preamp tubes. No effect.
- Checking voltages. They all look normal enough to me, but I'd appreciate another eye cast over these. This is my second amp build so I only know what I've read on the forum - particularly I don't really know what to expect from the PI. My B+ is a touch low because I've got a 200 ohm sag resistor in place for now (I might take this out - I kinda like how it feels, but it drops too much B+ from what I've read). Here are the voltages:


B+ (off the rectifier, after the sag resistor): 322 (without the sag resistor this is 355)
Screen supply: 306
PI plates supply: 270
Preamp tubes supply: 263

V1 (input stage of Normal + Wreck channels)
Pin 1 (Plate): 175
Pin 3 (Cathode): 1.30
Pin 6 (Plate): 169
Pin 8 (Cathode): 1.39

V2 (Wreck channel stages 2 and 3 - probably not relevant)
Pin 1 (Plate): 236 (this is a cold clipper stage)
Pin 3 (Cathode): 2.46
Pin 6 (Plate): 172
Pin 8 (Cathode): 1.35

V3 (PI)
Pin 1: 198
Pin 2: 50ish (forgot to take an accurate reading)
Pin 3: 71.2
Pin 6: 207 (note unbalanced PI so this is higher than pin 1)
Pin 7: 50ish
Pin 8: 71.2

V4 (power tube)
Pin 3 (Cathode): 10.5
Pin 7 (Plate): 319
Pin 9 (Screen): 299

V5 (power tube)
Pin 3 (Cathode): 10.5
Pin 7 (Plate): 320
Pin 9 (Screen): 298

Could this kind of thing be caused by a bad solder joint? I've reflowed most of them, and have a better soldering iron coming tomorrow.

If I can't find a problem, my current solution would be to increase the plate resistor to ~220k for more gain, then add a high pass filter / lower some coupling caps. But I'd rather work out why this is happening and do a proper job :P

Obligatory gutshot:
IMG_4557.JPG
https://drive.google.com/file/d/1hOF72_ ... sp=sharing

Any help is appreciated!

Regards,
Simon
You do not have the required permissions to view the files attached to this post.
0 x

geoff 1965
Verbose poster
Verbose poster
Posts: 1118
Joined: Mon 01/04/16 3:52 pm

Re: 18W TRex build - dark sounding

Post by geoff 1965 »

Hello Simon,
I had the ampmaker PT in my lite2 amp initially and only swapped it out to use a ez81 and the ampmaker OT’s are good transformers so I would’nt worry about those.if you swap the 200ohm sag for a 100ohm you will still get nice sag and only drop 10 volts off your B+.
Usually you have a 10nf into the phase inverter on the normal channel and yours is 22 so that could add to the extra bass also if you swap the 10nf from the tone pot to ground for a 22nf that will cut some bass.
0 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

Practice with your new soldering iron when you receive it, and watch the soldering videos linked in my signature. There are a lot of questionable solder joints in the photo.

I would also suggest swapping out the sag resistor for the 100 ohm to get your B+ where you want it (345VDC +/- 5V). Too low of a B+ will give you a muddier sound.

The other thing to do is highlighter verify every component connection and value using the schematic.

Question: what is the wattage rating for those metal film resistors? Those look like 1/4 watt resistors to me, and you should generally be using at least 1 watt resistors, with voltage ratings well above what any of them will see. I use CMF 1 watt metal film resistors.

Thanks,
Josh
0 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

Re: 18W TRex build - dark sounding

Post by SiRaff »

Wow, thanks for some very fast responses!

Re the caps, thanks for the tip. It would be a really convenient solution if it just involved swapping the two caps over (I'll check voltage rating on the 10n before I do this). I also have a spare 5n so will experiment. I'd still love to understand why it's so bassy, but it might be one of those things I just fix and never know why.

Sounds like you're both saying the same thing about the sag resistor, which is encouraging, because it would mean keeping the feel of the sag R without the large drop in B+. I'll get hold of a 100R and try that.

Agreed about the solder joints. I was quite embarrassed to post the photo actually. There are some horrible joints in there. I'm going to try and wick away the excess solder and do a better job. Will watch your video Josh, thank you!

The resistors are 0.6W even though they look like 1/4W. I hope this is good enough - it's too late to swap them all now so it is what it is!

Out of interest, a question on layout. I doubt I'll build another amp for a while, but I'd love to learn. I'm not that satisfied with the flying leads to the valve bases - crossing each other at all sorts of angles, some going directly over the valve bases to pins on the far side. How bad (scale of 1-10) is this practice? I'm getting almost no hum, and no oscillations, so empirically it's fine, but I'd love to know how to do this better.

Is the "whoosh" on the Wreck channel a normal part of the EL84 character? It's not a fizzing/buzzing so I don't think it's unwanted distortion. More like a smooth, long decay effect.
1 x

geoff 1965
Verbose poster
Verbose poster
Posts: 1118
Joined: Mon 01/04/16 3:52 pm

Re: 18W TRex build - dark sounding

Post by geoff 1965 »

No need to be embarrassed we all start somewhere,and if you’ve no hum or oscillations then that’s good you can work on your lead dress as you go.
swapping the 100R sag in will bring your voltages into the ball park for a start,also your pin3 voltage on the el84’s could come up a little,a 180R should get you nearer the 12 volts target.you want at least 5W wirewound for the bias.
good luck
1 x

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

Re: 18W TRex build - dark sounding

Post by SiRaff »

Thanks very much both of you. Re the plate resistors, is it the DC voltage or the (max) AC voltage across them that we are most concerned with? If it's DC, the voltage drop across the plates is more like 100V (supply node voltage minus voltage measured at plate). This would significantly drop the power going through them because P=V^2*R - so the DC power dissipation drops to 0.1W. This fits with some schematics I've seen where resistors are 1/2W unless stated otherwise. This seems consistent to me but have I missed something?

Under AC conditions, you might get the full ~270V dropped across the plate resistors, but not consistently - I'm wading out of my depth here but I think it introduces a factor of 1/2 (both RMS power and RMS current are 1/sqrt(2) of their peaks). Which would mean dividing your figure by 2 (power of 0.3645W, so 0.6W might just about be OK?). I'm happy to replace them but just interested in where the convention of "1/2W unless stated otherwise" comes from!

Very interesting about the PI actually being balanced. Now you mention it, I've seen this somewhere (maybe Merlin's book - the missus still doesn't understand why this is my choice for bedtime reading. She has no idea what she's missing. :D ). I'll try swapping out the 82k for 100k.
0 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

SiRaff wrote:
Wed 05/05/21 9:57 am
Thanks very much both of you. Re the plate resistors, is it the DC voltage or the (max) AC voltage across them that we are most concerned with? If it's DC, the voltage drop across the plates is more like 100V (supply node voltage minus voltage measured at plate). This would significantly drop the power going through them because P=V^2*R - so the DC power dissipation drops to 0.1W. This fits with some schematics I've seen where resistors are 1/2W unless stated otherwise. This seems consistent to me but have I missed something?
The question is of the "voltage across" not the "voltage drop." If there is voltage feeding a resistor, the voltage drop is a percentage of the voltage across it, but not all of the voltage across it. All of the voltage feeding the resistor goes through the resistor; the voltage drop is expelled as heat, and the rest of the voltage goes through. Think of a small dam that lets a specific amount of water through. The dam feels the pressure from all of the water, though not all the water comes out the other side.

In some circuits, and some areas of these circuits, 1/2W is sufficient. For the amp docs I write, they're generally all 1W unless otherwise noted.

Thanks,
Josh
0 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

SiRaff wrote:
Wed 05/05/21 9:57 am

Under AC conditions, you might get the full ~270V dropped across the plate resistors, but not consistently - I'm wading out of my depth here but I think it introduces a factor of 1/2 (both RMS power and RMS current are 1/sqrt(2) of their peaks). Which would mean dividing your figure by 2 (power of 0.3645W, so 0.6W might just about be OK?). I'm happy to replace them but just interested in where the convention of "1/2W unless stated otherwise" comes from!
Power dissipation is I^2 X R. This is a Class A stage, so average current is always the same. It might change temporarily when the stage is ovedriven, but that generally causes anode current to temporarily decrease as bias moves closer to cutoff. The bottom line is that DC parameters are used to calculate the resistor's dissipation. AC doesn't enter into the equation.

Jack

Edit: So that's the rationale for determining dissipation of the anode resistor in a Class A amplifier. In your specific case, because we know the voltages, E^2 / R can be used. The 100K resistor has 72V across it. It's dissipating approximately 50mW, or one-tenth of its rating if it's a 1/2W resistor. That would be a good safety margin in my book. The 82K resistor calculates to roughly the same.
0 x

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

JMPGuitars wrote:
Wed 05/05/21 6:16 am
Ohm's law tell us that for your PI plate resistor is being fed 270VDC to a 100K resistor, your minimum rating is .729 Watts, and you never want to be at the minimum, so 1W should be used.
Josh, this is not correct. The only voltage that matters is the differential across the resistor. That's how dissipation is calculated.
If you're also using an 82K, your power requirements are even higher. Don't use an 82K, it will sound better with 2 100K resistors, which is actually how you unbalance the PI. 82K + 100K plate resistors is an attempt to balance the PI. Balance doesn't mean equal, it means within a margin of error.
This will be true if we defeat the negative feedback. Otherwise, feedback will change the relative signal current through each side of the LTP, partially forcing its AC output voltage back into balance. There will be an increase in distortion as a matter of closed- vs. open-loop gain (the feedback isn't infinitely large), but the most significant difference in operation will actually be the loss of headroom that occurs in the unbalanced scenario. This amp has a switch for NFB, so the difference in tone with an unbalanced PI can be easily compared, and I would expect it to be significant with NFB off.
You should at the very least replace all your plate resistors.
I often use overrated components, and I seldom argue against this philosophy. Still, I think it's important to understand when there is or isn't a need to do so. In this case, undoing the previous work doesn't appear to serve a useful purpose.

Jack
0 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

TriodeLuvr wrote:
Wed 05/05/21 12:42 pm
JMPGuitars wrote:
Wed 05/05/21 6:16 am
Ohm's law tell us that for your PI plate resistor is being fed 270VDC to a 100K resistor, your minimum rating is .729 Watts, and you never want to be at the minimum, so 1W should be used.
Josh, this is not correct. The only voltage that matters is the differential across the resistor. That's how dissipation is calculated.
If that's not correct, then you should write an email to every Ohm's calculator on line, including: https://ohmslawcalculator.com/ohms-law-calculator

I could be wrong, I learned this by reading the internet. But it doesn't make sense to only consider the differential. The differential may be the the heat generated, but it still has to tolerate all of the juice passing through it.
If you're also using an 82K, your power requirements are even higher. Don't use an 82K, it will sound better with 2 100K resistors, which is actually how you unbalance the PI. 82K + 100K plate resistors is an attempt to balance the PI. Balance doesn't mean equal, it means within a margin of error.
This will be true if we defeat the negative feedback. Otherwise, feedback will change the relative signal current through each side of the LTP, partially forcing its AC output voltage back into balance. There will be an increase in distortion as a matter of closed- vs. open-loop gain (the feedback isn't infinitely large), but the most significant difference in operation will actually be the loss of headroom that occurs in the unbalanced scenario. This amp has a switch for NFB, so the difference in tone with an unbalanced PI can be easily compared, and I would expect it to be significant with NFB off.
Defeating the NFB is the only correct answer. 😉 NFB is terrible in 18watters, and sucks out the life and touch responsiveness.


Thanks,
Josh
0 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

JMPGuitars wrote:
Wed 05/05/21 2:57 pm

If that's not correct, then you should write an email to every Ohm's calculator on line, including: https://ohmslawcalculator.com/ohms-law-calculator
The calculator works, but you're not applying it correctly. If you enter 70V (voltage across the resistor) and 100,000 ohms, it returns a dissipation value of 0.049W. That's the right answer.
I could be wrong, I learned this by reading the internet. But it doesn't make sense to only consider the differential. The differential may be the the heat generated, but it still has to tolerate all of the juice passing through it.
I don't know what "juice passing through it" means. If you mean current, that's strictly a function of the resistance and the voltage differential. I=E/R The heat dissipated by the resistor is in no way related to its elevation above or below ground. Only the voltage across the resistor is relevant.
Defeating the NFB is the only correct answer. 😉 NFB is terrible in 18watters, and sucks out the life and touch responsiveness.
I realize NFB has been stripped from your designs (and the amp under discussion). The only reason I mentioned it is because it's an integral part of the Marshall presence control. It's difficult for me to ignore it, and I thought the concept regarding its relationship to PI balance might be of interest. :)

Jack
0 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

TriodeLuvr wrote:
Wed 05/05/21 3:52 pm
The calculator works, but you're not applying it correctly. If you enter 70V (voltage across the resistor) and 100,000 ohms, it returns a dissipation value of 0.049W. That's the right answer.
Okay, I was wrong. But why doesn't the total voltage matter? Is it because the resistor doesn't care because it isn't being released as heat?

I think it's the term "voltage across the resistor" I don't like, it's ambiguous. All the voltage goes across the resistor, not just the voltage dropped.

Thanks,
Josh
0 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

I'm not sure exactly how to explain this further, so let me just throw out an example to see if it helps. Imagine there's a 1.5 ohm resistor on your bench. You connect a 1.5V battery across it. You now have 1.5V across the resistor, and Ohms Law says this will create 1.5 watts of heat (E^2 / R). Now connect one side of the resistor/battery combination to the high voltage output of your bench supply, say +400V. Don't connect the other side to anything at all. The voltage across the resistor (aka the differential) doesn't change; it's still 1.5V. The dissipation of the resistor doesn't change, either. It's still 1.5W.

As a final example, let's connect one end of the resistor/battery combination to your +400V bench supply and connect the other end to the anode of a vacuum tube that's biased completely off. Even though the resistor has now been elevated by the +400V supply, it still has the same 1.5V across it and still dissipates 1.5W. The fact that it has been elevated to +400V makes no difference.

Jack
1 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

TriodeLuvr wrote:
Thu 05/06/21 11:24 am
I'm not sure exactly how to explain this further, so let me just throw out an example to see if it helps. Imagine there's a 1.5 ohm resistor on your bench. You connect a 1.5V battery across it. You now have 1.5V across the resistor, and Ohms Law says this will create 1.5 watts of heat (E^2 / R). Now connect one side of the resistor/battery combination to the high voltage output of your bench supply, say +400V. Don't connect the other side to anything at all. The voltage across the resistor (aka the differential) doesn't change; it's still 1.5V. The dissipation of the resistor doesn't change, either. It's still 1.5W.

As a final example, let's connect one end of the resistor/battery combination to your +400V bench supply and connect the other end to the anode of a vacuum tube that's biased completely off. Even though the resistor has now been elevated by the +400V supply, it still has the same 1.5V across it and still dissipates 1.5W. The fact that it has been elevated to +400V makes no difference.

Jack
That does make sense, thanks. I think that was also part of my misunderstanding- what I read was mostly based on a battery analogy.
0 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

Josh, you have no idea how many engineers I've worked with (or after they left) who got this wrong. Sometimes their calculations went the other way, and all the bias resistors on dozens of ICs burned up after the product was in the field six months or so.

Just to throw out one more thing (I know that's what you were hoping for), using a resistor that's too large can have consequences. Sometimes it's a good thing to have the resistor burn and open when the component it feeds fails and shorts. Otherwise, other components in series with the resistor, such as other dropping resistors, power supply diodes, etc.) will also fail. That's how solid state works - one transistor shorts, and a whole raft of them send up smoke signals. :lol:

Jack
1 x

User avatar
JMPGuitars
Super Duper Admin
Super Duper Admin
Posts: 3954
Joined: Tue 09/18/12 8:00 pm
Location: South Central, MA
Contact:

Re: 18W TRex build - dark sounding

Post by JMPGuitars »

TriodeLuvr wrote:
Thu 05/06/21 12:38 pm
Josh, you have no idea how many engineers I've worked with (or after they left) who got this wrong. Sometimes their calculations went the other way, and all the bias resistors on dozens of ICs burned up after the product was in the field six months or so.

Just to throw out one more thing (I know that's what you were hoping for), using a resistor that's too large can have consequences. Sometimes it's a good thing to have the resistor burn and open when the component it feeds fails and shorts. Otherwise, other components in series with the resistor, such as other dropping resistors, power supply diodes, etc.) will also fail. That's how solid state works - one transistor shorts, and a whole raft of them send up smoke signals. :lol:

Jack
Thanks. Yes, I'm aware of the second part, but in guitar amps it's almost always better to be overrated than underrated. Though I think I saw you or somebody else discuss this in one of our screen resistor conversations? Maybe an underrated screen resistor blew, and saved something as a result. I forget.

Thanks,
Josh
2 x
'I installed a skylight in my apartment yesterday... The people who live above me are furious.' - Steven Wright

Modern Ground Schemes
Soldering Technique
B+ Voltage Reduction
Amplifier Tools & Parts Info


Web Design: DolceVittoria.com
Guitars / Amps / Effects: JMPGuitars.com
(anti)Social: Facebook · Instagram
Items for Sale

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

Re: 18W TRex build - dark sounding

Post by SiRaff »

I've made a few tweaks based on advice on this thread and wanted to circle back with some findings. I have now finished work on the circuit (and don't want to revisit it unless something goes wrong!) and all that remains is to order a faceplate and make the whole thing look the part. It's currently just a blank chassis with lots of gummy glue residue on the front (yum), which is a hangover from its days as a humble MG50DFX. Overall I'm really delighted with the amp and it has some great sounds. Funnily enough, my favourite sound is a Strat neck pickup through the Wreck channel at around 4-5 on the Volume (gritty, but not distortion territory). It's lovely and articulate, with a beautiful compression when you dig in. I'll post sound clips soon.

The original issue was a dark sounding Normal channel with not all that much gain. This has improved with a couple of tweaks, although to my ears its still reasonably dark and I would always run the Tone control at full. The tweaks I've ended up making are:
- Swapped the 10n in the Tone circuit for a 22n
- Swapped the 22n coupling cap on the Normal channel for a 10n to let less bass through
- Unbalanced the PI by replacing the 82k with a 100k plate resistor. It does sound better this way, and the removal of NFB is indeed more noticeable.
- Turned the amp up! I've been testing the amp with an attenuator down to bedroom levels. The attenuator is a simple homemade affair, it does have a treble peaking cap but even so...the amp sounds brighter at higher volumes. I'm sure those Fletcher-Munson curves have some role to play in this.

I've made a small number of other tweaks based on feedback/advice from this thread, and general tinkering.
- Increased the cathode bypass cap in the PA to 2200uf/35V. This did tighten up the bass a lot, you can really hear it on palm muted notes.
- Added another sag resistor in parallel to drop it down to ~110 ohms (I was too lazy to order one).
- Added Merlin's tube-buffered FX loop. I know a loop in an 18W is pointless, with the OD being generated in the PA. But I had a spare tube socket, I was curious, and the loop is extremely transparent. It doesn't affect the tone or feel of the amp at all. One benefit for me was that I could add a cheap and nasty footswitchable gain boost, because the loop includes an extra gain stage after the Return jack. I simply added a voltage divider after the cathode follower (before the Send), and then wired the footswitch so that it parallels a resistor into the lower half of the divider (thus reducing gain when engaged). Voila, a safe lead boost with minimal effort. It is a little subtle (I chose values to give ~30% boost), and there's no LED to say when it's on. But it's there and it adds variety.
- Sorted out some of those bad solder joints. I got a new 65W iron. My soldering technique is still horrible, but it's the best I can do for now. It works now and if it breaks later, I'll fix it up.

I can't think of any other tweaks I've done. I'll edit this if anything comes to mind. If anyone is curious about the TRex (I guess there haven't been that many builds), happy to give more details!
2 x

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

Re: 18W TRex build - dark sounding

Post by SiRaff »

I forgot to mention I’ve also added VVR (I used the AmpMaker kit). It was a bit of an experiment, overall I like it. I scaled the whole amp - it was too much faff to add a separate power chain for the preamp, and scaling only PI & PA seems to come with its own issues from what I’ve heard (eg the need for a PPIMV).

At lower settings (loud TV kind of levels), it’s very usable. The overall gain definitely reduces compared to running at full B+, but as long as the VVR isn’t set too low, the amp is still responsive and the tone is good. It seems that the way to see VVR is that you’re not getting exactly the same amp at lower volume (the way our ears work, that’s impossible anyway). You’re getting a close cousin, which hopefully will behave in a similar way and have usable tones of its own.

At the very lowest settings, VVRing the whole amp seems to be disappointing. The amp becomes dull and not very responsive. But we’re talking conversation levels here. A better solution might be medium VVR setting plus a little attenuation. If I had the appetite, I’d go back in and install a half power switch (or maybe 1/10th power? Which would halve the volume?). Anyway, I’d be interested to hear how these observations compare to others who have tried this!
2 x

SiRaff
Unrated
Posts: 6
Joined: Sun 03/19/17 2:53 am

Re: 18W TRex build - dark sounding

Post by SiRaff »

I've discovered that I have the "buzz" problem that Paul Ruby discovered (the one described in this thread: viewtopic.php?t=18869).

His suggested fix with zener diodes seems to have cured it completely, so I'm posting this in case it helps others who might have the same issue.

When I was investigating, I struggled to find sound clips of the infamous buzz sound so it was hard to work out if this was really the problem. It's only by trying the fix (admittedly very cheap and easy to do) that I confirmed it. So, I've recorded a sound sample with (first half) and without (second half) the fix.

It's a little hard to hear the difference on the clip, probably because it was recorded on a phone. You can hear the buzz sound most clearly when I hit the individual low G notes in the second half of the clip (the sequence was: open G chord, G power chords, individual G notes, D power chords, high G note). It sounds kinda like a phaser, a high pitched "peeeesssshhh" sound that gradually decreases in pitch as the notes decay.
You do not have the required permissions to view the files attached to this post.
2 x

User avatar
TriodeLuvr
Frequent poster
Frequent poster
Posts: 616
Joined: Tue 10/16/18 11:19 am

Re: 18W TRex build - dark sounding

Post by TriodeLuvr »

SiRaff wrote:
Wed 06/09/21 3:18 pm
I've discovered that I have the "buzz" problem that Paul Ruby discovered (the one described in this thread: viewtopic.php?t=18869).

His suggested fix with zener diodes seems to have cured it completely, so I'm posting this in case it helps others who might have the same issue.

When I was investigating, I struggled to find sound clips of the infamous buzz sound so it was hard to work out if this was really the problem. It's only by trying the fix (admittedly very cheap and easy to do) that I confirmed it. So, I've recorded a sound sample with (first half) and without (second half) the fix.

It's a little hard to hear the difference on the clip, probably because it was recorded on a phone. You can hear the buzz sound most clearly when I hit the individual low G notes in the second half of the clip (the sequence was: open G chord, G power chords, individual G notes, D power chords, high G note). It sounds kinda like a phaser, a high pitched "peeeesssshhh" sound that gradually decreases in pitch as the notes decay.
One of the scope displays in Paul Ruby's PDF shows this.

http://www.18watt.com/storage/18-watter ... fo_311.pdf

Scroll down to "Speaker Load Comparison," and the lower trace tells the tale. Both the positive- and negative-going halves of the waveform are clipped, and the clipped areas clearly contain another (higher) frequency. It appears to be roughly twice the original, which indicates rectification is occurring. This secondary frequency will mix back with the fundamental to create a slew of higher order harmonics. The harder the amplifier is pushed, the greater the amplitude of these products, and the more audible the higher frequencies become. This is almost certainly the reason for the apparent pitch decrease as the notes decay.

I haven't had time to study Ruby's mod in depth. My only question, really, is how it counteracts the increase in bias that occurs simply due to increased current through the cathode bypass network. That in itself is sufficient to create crossover distortion in AB1 amplifiers, and it won't be affected by his modification. A little time with the simulator will probably reveal what's happening.

Anyway, congrats on your work and the success of all the mods!

Jack
1 x

Post Reply