My step into the single-ended world - DIY 3W mono blocks

[Blasty]

Hello everyone, I hope this post finds you well.

This might get lengthy and I still won't hit every point I wanted, but here's an initial writeup.

Feel free to skip and look at the pictures/schematic that I have attached.

Earlier this year I decided it was time for another electronics project.

I have built my own headphone amp and phono preamp, but I have never built a power amp. I have been using a Magnavox 185 converted to a 175 since I was around 16, it is really the only tube sound on loudspeakers that I am familiar with. A single-ended amplifier seemed like a good project and would be something different to try. Or how about two single ended mono amplifiers! This is what I came up with.



The output stage is based on this great article by Matt Renaud: http://diyaudioprojects.com/Technical/6V6-SE-UL-Bias-Optimization/

The two-stage input is the "Low-Distortion Gain Block" designed by Max Robinson: http://www.angelfire.com/electronic/funwithtubes/Amp-Gain_Block.html

Iron is by Edcor, blank chassis, cage, and bottom cover by Hammond, faceplate designed by myself and manufactured by Front Panel Express.

Since I intended to feed the amplifier with either my phono preamp (~0.7V out) or Schiit Modi 2U DAC (1.5V out), both through a passive preamp, I wanted to be able to set the sensitivity of the amplifier to accommodate either. In Max's circuit, the gain is determined by the resistor to ground on the first stage. SW2 on my schematic allows this resistor to be changed.

After building my amplifiers, I wouldn't recommend doing it this way. In my SPICE simulation it does what it's supposed to. It And it does in real life, except with a BANG! through the speakers when the switch is thrown. Best to turn the amplifier off when operating this switch.

In the future I may change it such that the amplifier is permanently configured for 0.7V sensitivity, and use SW2 to control an attenuating resistance instead. Let's move on.

With a shorting plug installed and my ear against the speaker, there is only the very faintest shot noise heard in the tweeter. On any other driver there is only silence. This is on Klipsch La Scalas having a sensitivity of 104 db/W, 1m. On the oscilloscope, I see less than 2mV of noise.

I attribute the low noise floor to several components. I might spend an inordinate amount of this post on these items, but since hum and noise are a frequent issue faced by the DIY'er it seems appropriate to share:

• - All AC and the rectification is confined to approx. 1/3 of the chassis, located opposite the sensitive input stage.
• - The IEC AC socket contains a built-in EMI filter network.
• - DC filtering is generous, The circuit is a CLCRC arrangement using a 15H choke followed by 100uF capacitors.
• - Signal ground and earth ground are separate, except for an RC hum blocking network between them. The resistor also provides a return path in case of a fault to chassis on the secondary (isolated) side of the power transformer, which is why it is a wire-wound power type.
• - The reservoir capacitor negative lead and secondary center tap are tied to one end of a straight ground bus. Next are the returns from the power stage, and finally the input stage. Each stage's grounds connect near the negative lead of the associated filter capacitor.
• - Heater wires outside of the power supply area are twisted and routed out of the way of sensitive components. In this layout I honestly think I could have got away with not twisting them.
• - Heaters are biased to the same potential as the output tube cathode, eliminating the unwanted electron flow and associated noise that can occur between the heaters and cathodes.
• - The steel chassis provides both electrical and magnetic shielding. (I have always built on aluminum before).
• - Resistors are metal film types. Preferences aside, they should contribute less noise than carbon types.
• - Tube sockets are ceramic with silver plated contacts for reliable connections. Dealing with noise from socket contacts is a pain so I feel it was worth it to get a little bit fancy here.

Now for listening. I can't pretend to know what I'm talking about, since my experience is almost entirely with my Magnavox. I've also only had these together for less than two days. I listened for a while on just one while I worked on the other, but that doesn't give a very accurate impression when I'm used to stereo. I honestly didn't know what to expect from a single ended setup which is part of the fun of trying it. I will say that no veils were lifted, etc. Maybe just blown to the side a little bit . The room did not disappear, nor did the musicians come in, nor was I transported there... not entirely. Some music sounds very different on these. Good recordings, especially the few I have with real, unprocessed instruments can sound eerily lifelike. I think this is one of the appeals of single-ended. Mostly, all my music still sounds very good but now it's played on something that I built!

I have much more listening and tweaking around to do before I can report further on sound. Sorry if that's what you were waiting for. Guess you'll have to build these and figure it out yourself .

Seriously, I'm just going to take a break from typing this and come back later to contribute more information as needed.

I hope you all enjoy looking and learning from my DIY endeavor as much as I do yours.

 

[kcdonahue]

WOW

They look amazing... Wish I were close enough to drop by for a listen!

KC

 

[turnitdown]

That is some tight work. VERY nice.

 

[Blasty]

Thank you for the compliments

Apologies for the haphazard introductory post, how does one sum up months of planning and design? Maybe I should break out of my shell and do a video explaining them at some point.

It may have come off reading like the end result was no big deal... it wasn't meant to. I had no idea what to expect from these, and let's face it, the audio world is full of so many different opinions regarding different topologies that I felt the need to be careful what I say about the sound.

I will come clean and say I was a little underwhelmed when I finally got both completed and listened in Stereo. But I made a classic mistake of changing more than one thing in the system. I had connected these straight to the preamp output instead of through my equalizer that I normally use. Once I put the equalizer back in to tweak the sound to my liking, it became apparent that these do offer something that my Magnavox (which has been reworked plenty and sounds great) does not.

I will also add that I did not expect a "you are there" or "they are here" experience because my room simply isn't conducive to it. My den is a general purpose entertainment area and not just an audio room, and I'm fine with that. I am happy to report that the mono blocks still make a difference. How to describe it? I don't know, and I don't really care as long as they do what they do.

I am excited to revisit my music collection with these.

For those who may wonder, I don't operate a company called Windom Engineering. I just felt that since I was building these to hopefully outlast me, that a "brand" with my name on it was appropriate. It could be a mystery for the future owner to wonder about.

 

[I LIKE MUSIC]

Very nice work and write up.

Thanks for the pictures.

 

[BillWojo]

Very nice build, I like them. Maybe 50 years from now someone will find these and do a search on AK!
By all means, do a video, should be interesting.

BillWojo

 

[macaltec]

Very clean work.

 

[MickeySmile]

"like" big time!

 

[gadget73]

I can see why changing the cathode resistor value with the switch would cause a pop. Basically you'd be moving the cathode voltage away from the grid, exactly like if you'd pulsed a signal to the input rapidly. What you could do is use a variable resistor across a fixed resistor, and mix the two values to get the range you want. You'd still be changing the cathode voltage but it would be a lot less sudden, which should keep it from making a loud pop. You'd probably see the woofer "pump" as you moved the pot though.

Nice looking build.

 

[Blasty]

gadget73:

Thanks for that idea, I agree that would be the best way to accomplish a smooth switch if I must have control over the cathode resistance.

However, I'm thinking this was a clever "nice to have" idea that I hadn't seen elsewhere for a good reason. I never thought about the square step when running simulations comparing with two different resistances.

Since I'm in for those nice face plates already, I may leave it alone since it does its job when operated correctly. Otherwise I'll test feeding the input with a voltage divider and the switch across the top resistor. Luckily I can just insert a test version of the circuit in-line with the input jack before I commit to changing anything in the amps themselves.

When I put enough time and attention into a project I sometimes get confident enough to forgo a breadboard test with the mindset that some minor tweaking may be required; lesson learned this time. Yet if that is the only flaw that shows up in a project like this, I'm still very happy.

 

[gadget73]

Unrelated, but is that one of those Chrysler hemi-powered air raid sirens in your avatar?

 

[Blasty]

Yes, the Chrysler-Bell Victory Siren. It is the predecessor to the later version with the V8 Hemi. I forget the engine used in the earlier one. They are all but extinct. A few years ago I was aware of two in China that are still in service. One had been converted to use an electric motor, the other had some other engine in place of the original.

138dB at 100'

 

[gadget73]

DeSoto Firedome maybe?

yeah they were a fairly insane piece of gear, developed at a somewhat insane time in history. I guess all times are insane in their own way, but seriously why did anyone think hiding under a school desk would make you safe?

 

[Blasty]

Sorry to bump this back up, but wanted to give my impressions after listening to these for a while.

It may have read at first like I wasn't impressed with the sound. However, after giving MY EARS some break-in time (yeah, that's right!) I'm enjoying them more. I went into this not knowing what to expect, having never heard a proper single-ended amp. Proper being not the output stage of a table radio. From all my reading on the single-ended sweetness, even-ordered harmonics, etc. I think my expectations were for a more pronounced change.

I did notice right away that the midrange is a bit more forward than with the Magnavox PP amp and its global negative feedback to smooth things out. It was almost biting on certain notes, but damping some hard surfaces in the room (notably the glass coffee table -- which is temporary anyway) took care of most of that. It gave me back a bit of bass, too.

Aside, some might cringe at the thought of using an amp with a forward midrange on Klipsch K-400 horns, can you imagine? Sorry, I made my self laugh at the thought of someone chiming in to say, "Of course it sounds like an ice pick, you dummy!" But just damping hard surfaces helped a bunch and I'm back loving the sound.

As I peruse through my music collection, I am beginning to recognize what these amps do well. Some guitar strings have a very real quality that commands attention, where I hardly noticed before as they may have been just part of the background detail. Example, beginning at about 1:50 in "Secrets" by The Cure. I'm beginning to realize what a great sounding album (Seventeen Seconds) this is, not just another of my '80s discs.

I will echo what I've read and say that I'm finding vocals to also have a very convincing presence.

I'm sure I'll notice more as I get acquainted with these. Any time I notice little details that I didn't think much about before, I feel I'm on the right track.

High quality recordings are a treat to listen to. I'm playing some Porcupine Tree right now. Middle-of-the-road stuff still sounds great. Much of my collection consists of well-used (but not abused) records and many CDs I picked up because I like the music. I haven't intentionally cherry picked any of it for recording quality. However, I'm to the point with my system where the recording quality is more noticeable than it used to be. What goes in is what comes out!

I think I've built some solid amplifiers that I'll enjoy for a long time.

 

[Blasty]

I'm resurrecting this thread for anyone who is considering trying this circuit. The fact that the input level switch was poorly implemented kept bothering me, so I have changed how it operates. If implemented as shown in the attached schematic, there is no *BANG* when the switch is thrown.

The attached pictures show a before (REV_A) and after (REV_B) of the input area. I've also attached a PDF file of the revised schematic.

Instead of changing the gain of an active stage, The gain of the amplifier is fixed for a 0.7V input. The input level required for full output is now determined by a passive attenuator before the first stage. To drive the amplifier to full output with ~0.7V in, the switch bypasses the attenuator so that the amplifier is driven directly at the input capacitor. For ~1.5V to reach full output, the switch is opened to put the series resistor of the attenuator in-circuit.

If I were to do this over, I would probably omit this feature entirely. It is nice to have, but not necessary. Keeping it fulfills the original intent of the design and keeps me from having to spring for a new set of faceplates.

The concept was simulated, then tested with an external arrangement before re-working the amplifiers themselves.

BEFORE:


AFTER:

 

[Tinkerbelle]

Thank you for providing the latest updates on this project - this should be useful for anyone wishing to try this or similar.

 

[BinaryMike]

Thanks again. I didn't appreciate that clever touch with increasing the input impedance on the first go-round. Do you hear a difference in HF response between the two gain settings?

 

[multichamp12]

Should both sides of your input jack wire shield be grounded? I have been under the impression it's common practice to only connect the shield on the input jack end only and run a ground wire from the input jack to the ground bus.

 

[Blasty]

I never heard a difference in HF response with either input level selector configuration (before or after rework), and from the simulations I didn't expect to. I don't have accurate measurements of actual frequency response at this time, just some cursory checks on my somewhat cheap oscilloscope to make sure all appeared well. Eventually I should work on a plot, I have fun putting together the documentation for these projects.

The input jack is only grounded at the end of the square bus wire visible in the rework photos. It is installed into the chassis with a plastic shoulder washer to keep it isolated.