[Xcalibur255]

Well a number of factors, but the big one is that a pentode amp would just be too much output voltage for headphones.

You would end up having to be very careful with the volume control, or you would have to have some voltage divider network on the output to ensure that you wouldn't blow your headphones.

Negative feedback reduces gain right? So a ton of feedback would fix the problem but probably kill the sound at the same time. My guess was 180 degrees off the mark.

 

[L0rdGwyn]

Negative feedback reduces gain right? So a ton of feedback would fix the problem but probably kill the sound at the same time. My guess was 180 degrees off the mark.

Negative feedback does reduce gain. It is often used with pentode stages because 1) there is typically an excess of gain, 2) a high output impedance, 3) high distortion (relative to a triode). Adding negative feedback will linearize the stage and reduce the output impedance, the amount of "extra" gain in the stage will determine how much NFB you are able to apply, you can run into stability issues if you use an excess of NFB. In my experiments using local NFB in my 801A amplifier over the past several months, it is all about implementation. Local NFB can vastly improve the sound of an amplifier if done right. My initial experiments resulted in a third-harmonic dominant distortion spectrum, which resulted in a "tinny" and "thin" sound, not good. I've made changes to achieve a second-harmonic dominant distortion spectrum with much better results sonically.

Certain prototyped versions of this amplifier I can confidently say were the best I have ever heard FWIW. Some notable changes have been made, most significantly changing the source follower buffer stage to a cathode follower and the loading / biasing scheme of the pentode input, there are now very few transistors involved. We'll see how it turns out! But count me as one who is not afraid to use NFB, it just needs to be done right.

 

[Tjj226 Angel]

Negative feedback reduces gain right? So a ton of feedback would fix the problem but probably kill the sound at the same time. My guess was 180 degrees off the mark.

It does, but remember that at 100% plate to screen feedback would be a triode strapped pentode and the gain would still be a factor of 30.

Even in triode mode, the amp will have a pretty hot voltage output. Let's assume that he will use a 2.5K to 32 ohm transformer for headphones. That would mean that the voltage step down ratio is about 9:1. By the time you add in losses and everything else we can say that the real world step down is 10:1 for simplicity sake.

When you consider that lordgwyn will be using a 1:2 step up input transformer in combination with the triode strapped E55L, you end up with an effective gain of about 60. Divide that by the step down ratio of 10 and you get a theoretical gain of about 6 for the overall amplifier. With a 2V input you are getting 12v out.

That is a pretty hot output for a headphone amp, but it isn't too crazy. High impedance headphones can benefit from such a high voltage output and 12V is manageable with a volume control. So considering that the amp is triode mode is already more than enough, imagine what an SEP amp would look like with less than 100% NFB.

Now he could design it such that the amp would be in triode mode for 32 ohms or 8 ohms, and then have an SEP mode that is locked to 8 ohms. That way you can just run the headphones off the 8 ohm tap and let the transformer deal with the excess gain and noise floor from the pentode operation.

 

[L0rdGwyn]

When are you going to wind your own OPT @Tjj226 Angel ? I need to live vicariously through you.

 

[Tjj226 Angel]

When are you going to wind your own OPT @Tjj226 Angel ? I need to live vicariously through you.

By the end of march, I should be more free. The goal is to build a CNC winding machine first. I would like to build one similar to the one in this video



The thing I like about this machine vs something you can buy off some chinese company is that the wire tensioner should be able to accept almost any gauge of wire. It's also relatively small compared to the size of the transformer you can make. Once that is all done, winding a transformer is simply a matter of doing the math for whatever transformer I want. The next step would be to wind field coil motors and make my own speaker drivers. One of my long held pasion projects is to take this speaker here...



...and miniaturize it for desktop use. Western electric 555 compression drivers are what makes that system really shine, but they cost more than my two cars, computers, and monthly rent combined. So I have to make it from scratch.

 

[L0rdGwyn]

By the end of march, I should be more free. The goal is to build a CNC winding machine first. I would like to build one similar to the one in this video



The thing I like about this machine vs something you can buy off some chinese company is that the wire tensioner should be able to accept almost any gauge of wire. It's also relatively small compared to the size of the transformer you can make. Once that is all done, winding a transformer is simply a matter of doing the math for whatever transformer I want. The next step would be to wind field coil motors and make my own speaker drivers. One of my long held pasion projects is to take this speaker here...



...and miniaturize it for desktop use. Western electric 555 compression drivers are what makes that system really shine, but they cost more than my two cars, computers, and monthly rent combined. So I have to make it from scratch.

To the naked eye that transformer winder looks very well-engineered, would be an incredible resource to have. Amazing speakers, I'm sure there would be high demand for a miniaturized version. Sounds like the makings of a profitable albeit niche business venture

 

[bcowen]

By the end of march, I should be more free. The goal is to build a CNC winding machine first. I would like to build one similar to the one in this video



The thing I like about this machine vs something you can buy off some chinese company is that the wire tensioner should be able to accept almost any gauge of wire. It's also relatively small compared to the size of the transformer you can make. Once that is all done, winding a transformer is simply a matter of doing the math for whatever transformer I want. The next step would be to wind field coil motors and make my own speaker drivers. One of my long held pasion projects is to take this speaker here...



...and miniaturize it for desktop use. Western electric 555 compression drivers are what makes that system really shine, but they cost more than my two cars, computers, and monthly rent combined. So I have to make it from scratch.

That winding machine is way cool. Couldn't see it in the video, but I assume there's some sort of counter or display that tracks the revolutions (which I also assume need to be precise)?

 

[Tjj226 Angel]

To the naked eye that transformer winder looks very well-engineered, would be an incredible resource to have. Amazing speakers, I'm sure there would be high demand for a miniaturized version. Sounds like the makings of a profitable albeit niche business venture

It is definitely a sight to behold. He has a couple videos on his channel where he shows more angles and features. The real takeaway is the wire tensioner. Everything else is fairly straightforward. To give some perspective, some wire tensioners being sold for transformer winders are friction based devices. They actually have a drag wheel that pinches your wire. This can work, but you risk damaging the insulation layer of the magnet wire.

The better wire tensioners will work by providing feedback to a motor that will let wire go from the spool at a very particular rate, however, they cost several hundred if not thousands. This guy's feedback mechanism is that little cart connected to the spring with a piece of linear tape to the side to monitor the position of the cart. This is actually how a printer works and most of these parts can be sourced from a scrap printer to make the cost of the project a lot cheaper while still maintaining a very high degree of accuracy.

It's pretty freaking ingenious. His design can also be made for a lot cheaper too. I get the feeling that his machine is built the way it is because he had a lot of those parts on hand. If you ditch a lot of the gears and belts by making all your motors direct drive, it should reduce the cost by quite a bit.

The ONLY modification I would love to make is replacing that spring on the cart with a load cell or a load cell and a spring. That way I can know the amount of tension I put on the wire and dial it in for best performance. Other than that, this machine is probably about as good as it gets.

I might end up sending you a few practice transformers for you to test considering you have more test equipment than I do : P

---------------------------------------

The speaker is a very different story. It's one thing to build that speaker. It is a completely different thing to figure out how to make it in such a way that it is easy to produce.

The real achievement will be making a smaller than average mid range horn. That KS-6368 horn on top of the speakers I linked there is nearly flat at 220hz. I want mine to be flat at 300hz, so I should be able to downsize it a bit. That horn is about 16x11 at the mouth, and I am really hoping I can get away with 12x9. If I have a 12 inch wide horn on top of a 12 inch wide woofer cabinet with a 10 inch woofer, it should work out to being just a bit bigger than the snells with better bass extension and much better mid clarity.

Definitely on the large side, but also doable for a large desk speaker. Or so I think.

 

[Tjj226 Angel]

That winding machine is way cool. Couldn't see it in the video, but I assume there's some sort of counter or display that tracks the revolutions (which I also assume need to be precise)?

It's completely computer controlled, so rather than a counter, the transformer bobbin is attached to a stepper motor and a computer tells the motor to step X number of times.

Theoretically with a bit more careful thought and engineering, you can set this up to be completely automated. If you really wanted to go crazy, you could have a multiple spindle setup and have some type of CNC taping machine in the mix so that you can slap on some bobbins and a butt load of wire at night and in the morning you would have your PSU, choke, input, output, and interstage transformers all waiting for you.

 

[whirlwind]

By the end of march, I should be more free. The goal is to build a CNC winding machine first. I would like to build one similar to the one in this video



The thing I like about this machine vs something you can buy off some chinese company is that the wire tensioner should be able to accept almost any gauge of wire. It's also relatively small compared to the size of the transformer you can make. Once that is all done, winding a transformer is simply a matter of doing the math for whatever transformer I want. The next step would be to wind field coil motors and make my own speaker drivers. One of my long held pasion projects is to take this speaker here...



...and miniaturize it for desktop use. Western electric 555 compression drivers are what makes that system really shine, but they cost more than my two cars, computers, and monthly rent combined. So I have to make it from scratch.

This is a great video of the winding machine....cool stuff. Thanks for posting

 

[L0rdGwyn]

It is definitely a sight to behold. He has a couple videos on his channel where he shows more angles and features. The real takeaway is the wire tensioner. Everything else is fairly straightforward. To give some perspective, some wire tensioners being sold for transformer winders are friction based devices. They actually have a drag wheel that pinches your wire. This can work, but you risk damaging the insulation layer of the magnet wire.

The better wire tensioners will work by providing feedback to a motor that will let wire go from the spool at a very particular rate, however, they cost several hundred if not thousands. This guy's feedback mechanism is that little cart connected to the spring with a piece of linear tape to the side to monitor the position of the cart. This is actually how a printer works and most of these parts can be sourced from a scrap printer to make the cost of the project a lot cheaper while still maintaining a very high degree of accuracy.

It's pretty freaking ingenious. His design can also be made for a lot cheaper too. I get the feeling that his machine is built the way it is because he had a lot of those parts on hand. If you ditch a lot of the gears and belts by making all your motors direct drive, it should reduce the cost by quite a bit.

The ONLY modification I would love to make is replacing that spring on the cart with a load cell or a load cell and a spring. That way I can know the amount of tension I put on the wire and dial it in for best performance. Other than that, this machine is probably about as good as it gets.

I might end up sending you a few practice transformers for you to test considering you have more test equipment than I do : P

---------------------------------------

The speaker is a very different story. It's one thing to build that speaker. It is a completely different thing to figure out how to make it in such a way that it is easy to produce.

The real achievement will be making a smaller than average mid range horn. That KS-6368 horn on top of the speakers I linked there is nearly flat at 220hz. I want mine to be flat at 300hz, so I should be able to downsize it a bit. That horn is about 16x11 at the mouth, and I am really hoping I can get away with 12x9. If I have a 12 inch wide horn on top of a 12 inch wide woofer cabinet with a 10 inch woofer, it should work out to being just a bit bigger than the snells with better bass extension and much better mid clarity.

Definitely on the large side, but also doable for a large desk speaker. Or so I think.

Looks great, will be excited to see where this goes, and happy to help out with testing if I have the necessary equipment need to make some upgrades come to think of it, particularly my function generator.

 

[L0rdGwyn]

I have been hard at work the past two days making preparations for the 801A amplifier.

Placed a mega parts order at Hificollective. Excited to show off what is going in, will post some pics of the parts pile when it arrives.

In the meantime, here is my little pile so far. Still need to stuff the Coleman regs.



I am finalizing one last Mouser order, then theoretically I will have everything but the mains transformer from Sowter...and some PCBs.

The bulk of the work the past two days has been designing power supply PCBs. This amp is fixed bias, will have a Maida regulated B+ and shunt regulated negative supply.

For the shunt regulated supply, I put together a raw DC board.



Then a pair of regulator boards that will be mounted in close proximity to where they will be used in circuit.



This is definitely not a "just build it" type of design, it will need a final prototype, so that is the last step before the layout and chassis can be worked on.

 

[m17xr2b]

I already have tungsten rectifiers so no reason to get them but I though they'd work well in your 842 amp.
https://www.ebay.com/itm/PAIR-NOS-I...ON-UX-216B-GLOBE-RECTIFIER-TUBES/303857004641

 

[L0rdGwyn]

I already have tungsten rectifiers so no reason to get them but I though they'd work well in your 842 amp.
https://www.ebay.com/itm/PAIR-NOS-I...ON-UX-216B-GLOBE-RECTIFIER-TUBES/303857004641

Very cool! Interesting rectifier. I'm pretty committed to solid state rectification, a pair of UX216B can't quite meet the DC current demands of the design (around 181mA) but thanks for the link. It is an 801A output rather than 842, would love to build with the 842, but that is one rare tube my headphone amplifier is an 841 build, with 842 I could complete the trifecta

 

[L0rdGwyn]

Found a much cooler vintage bakelite volume knob. The ones I purchased I worry are too small, might look / feel too dinky on the front panel of a 3.25" thick chassis.

The skirt of this one is 2" and I like the aesthetic 10/10 projected knob feel.