Hello everyone.
I have not started a new build thread in a while and this one may contain some material from previous threads.
I am finally getting around to building up what initially started as a ’76 2204 Marshall replica build. As I had done for the 1987 replica (https://ampgarage.com/forum/viewtopic.php?f=9&t=24683). Now it has morphed into a 2204 crossover build. This one will be less faithful to the vintage early second version 2204. It will be augmented with the modifications I have tested in recent years coming from my trusted current 2204.
For this project I drew the functional wiring schematic to adapt it to the component board. Before moving on to the construction stage, I wanted to get your expert opinions. What are your comments?
Here is the result, This is an updated version of the first layout initially published, notably featuring an improved approach to bias circuit design:
Alternate 2204 build
Moderators: pompeiisneaks, Colossal
Alternate 2204 build
You do not have the required permissions to view the files attached to this post.
Last edited by deeder on Thu Feb 12, 2026 1:19 pm, edited 3 times in total.
2203/2204-ish/Deluxe Reverb
Re: Alternate 2204 build
The circuit itself looks like it'd work well, though I have a few suggestions:
- Unless you happen to have a number of 2W 47K resistors on hand and non with lower power ratings, 2W is majorly overkill for the NFB resistor and I don't think will offer discernible performance advantage over 1W or 0.5W. Noise isn't really an issue this late in the signal chain, and even if your amp is producing voltage peaks of 30V (i.e., 112.5W output) a 47K resistor would dissipate just under 0.2W. So a 0.5W resistor would have a comfortable 2x margin of safety.
- I think you may run into noise issues with the lead dress proposed in your layout with the coupling cap associated with V1B being further toward the middle of the board than the V1A coupling cap. I'd suggest switching that order.
Do you plan on using the low gain input? I don't love having a peaking network and voltage divider feeding the input gain stage when that stage has a large, non-bypassed cathode resistor. That'll leave a lot on the table in terms of noise when using the low gain input since you're feeding a heavily attenuated signal into a stage that produces very little gain. Your schematic shows using switched Cliff-style jacks. If you do use that style of jack, you may want to try having a 20uF or larger cap bypassing V1A's cathode and leave the ground reference of that bypass cap connected to the sleeve switch associated with the high gain input. That way your high gain input retains the cold clipper since plugging into that jack opens the switch grounding V1A's cathode bypass cap, but your low gain input will have an improved signal-to-noise ratio since using the low gain input rather than the high gain input keeps the bypass cap's ground connection intact. See the cathode bypass network associated with the bright input in this schematic:
- Unless you happen to have a number of 2W 47K resistors on hand and non with lower power ratings, 2W is majorly overkill for the NFB resistor and I don't think will offer discernible performance advantage over 1W or 0.5W. Noise isn't really an issue this late in the signal chain, and even if your amp is producing voltage peaks of 30V (i.e., 112.5W output) a 47K resistor would dissipate just under 0.2W. So a 0.5W resistor would have a comfortable 2x margin of safety.
- I think you may run into noise issues with the lead dress proposed in your layout with the coupling cap associated with V1B being further toward the middle of the board than the V1A coupling cap. I'd suggest switching that order.
Do you plan on using the low gain input? I don't love having a peaking network and voltage divider feeding the input gain stage when that stage has a large, non-bypassed cathode resistor. That'll leave a lot on the table in terms of noise when using the low gain input since you're feeding a heavily attenuated signal into a stage that produces very little gain. Your schematic shows using switched Cliff-style jacks. If you do use that style of jack, you may want to try having a 20uF or larger cap bypassing V1A's cathode and leave the ground reference of that bypass cap connected to the sleeve switch associated with the high gain input. That way your high gain input retains the cold clipper since plugging into that jack opens the switch grounding V1A's cathode bypass cap, but your low gain input will have an improved signal-to-noise ratio since using the low gain input rather than the high gain input keeps the bypass cap's ground connection intact. See the cathode bypass network associated with the bright input in this schematic:
You do not have the required permissions to view the files attached to this post.
Re: Alternate 2204 build
cdmike, thank you very much for proofreading the layout; I really appreciate it. I'll look into the low-gain input topic further.
I also wanted to thank Sluckey for the Visio components; Steve Miller for his inspiration, Martin and Roe for their advice as well.
I made the turret board using Phenolic Paxoline P3: The chassis is from Valvestorm (thanks, Robert), and I had to modify it to drill the hole for the 50/50mF capacitor: The biggest challenge was building the headshell cabinet, and I'm quite happy with the result:
Regarding the components I'll be using for this project, I'll be drawing from my stash of parts that I've accumulated over time and have been wanting to use for a long time.
I also wanted to thank Sluckey for the Visio components; Steve Miller for his inspiration, Martin and Roe for their advice as well.
I made the turret board using Phenolic Paxoline P3: The chassis is from Valvestorm (thanks, Robert), and I had to modify it to drill the hole for the 50/50mF capacitor: The biggest challenge was building the headshell cabinet, and I'm quite happy with the result:
Regarding the components I'll be using for this project, I'll be drawing from my stash of parts that I've accumulated over time and have been wanting to use for a long time.
You do not have the required permissions to view the files attached to this post.
2203/2204-ish/Deluxe Reverb
-
Colossal
- Posts: 5332
- Joined: Sat Oct 20, 2007 9:04 pm
- Location: Moving through Kashmir
1 others liked this
Re: Alternate 2204 build
Very nice work. Will look forward to seeing the finished amp.
-
pdf64
- Posts: 3096
- Joined: Sat Mar 12, 2011 9:23 pm
- Location: Staffordshire, UK
- Contact:
3 others liked this
Re: Alternate 2204 build
Having the bias supply deenergised in standby doesn't seem a good idea?
https://www.justgiving.com/page/5-in-5-for-charlie This is my step son and his family. He is running 5 marathons in 5 days to support the research into STXBP1, the genetic condition my grandson Charlie has. Please consider supporting him!
Re: Alternate 2204 build
Thank you very much for this comment.
It seems that Marshall was using JMP era standby arrangement for 50 watters, which deenergises HT and bias supplies, resulting in a massive shock to the output valves and power supply when standby is flipped.
After quite a bit of research, I understood that the bias supply needs to be unswitched, like the heaters. The normal bias circuit design seems to take the bias voltage from the HT secondary before the standby switch, thus allowing the bias circuit to heat up in standby mode.
Therefore, I redesigned the bias circuit, which takes the high voltage directly from the secondary PT. And thus connects the 220k resistor:
I will use the initial Standby switch in SPST mode.
It seems that Marshall was using JMP era standby arrangement for 50 watters, which deenergises HT and bias supplies, resulting in a massive shock to the output valves and power supply when standby is flipped.
After quite a bit of research, I understood that the bias supply needs to be unswitched, like the heaters. The normal bias circuit design seems to take the bias voltage from the HT secondary before the standby switch, thus allowing the bias circuit to heat up in standby mode.
Therefore, I redesigned the bias circuit, which takes the high voltage directly from the secondary PT. And thus connects the 220k resistor:
I will use the initial Standby switch in SPST mode.
You do not have the required permissions to view the files attached to this post.
2203/2204-ish/Deluxe Reverb
-
pdf64
- Posts: 3096
- Joined: Sat Mar 12, 2011 9:23 pm
- Location: Staffordshire, UK
- Contact:
1 others liked this
Re: Alternate 2204 build
The standby in that Marshall image is fine, it's from the JCM era, and the bias supply stays energised whatever the standby setting.
https://www.justgiving.com/page/5-in-5-for-charlie This is my step son and his family. He is running 5 marathons in 5 days to support the research into STXBP1, the genetic condition my grandson Charlie has. Please consider supporting him!
Re: Alternate 2204 build
Thank you so much Pete.
I also remember you saying that it's better to switch the AC rather than DC; an AC arc will tend to self extinguish whereas DC will sustain itself, all else being equal. But with this arrangement I will switch the DC.
I also remember you saying that it's better to switch the AC rather than DC; an AC arc will tend to self extinguish whereas DC will sustain itself, all else being equal. But with this arrangement I will switch the DC.
2203/2204-ish/Deluxe Reverb