Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Feb Tue 25, 2020 10:47 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 7 posts ] 
Author Message
 Post subject: Solid state 1920's speaker amplifier revisited
PostPosted: Nov Fri 02, 2018 12:26 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
I had built the following solid state amp for driving 1920's speakers maybe 2-3 years ago and I rarely use it at the moment.

Attachment:
Amp 12.png
Amp 12.png [ 30.99 KiB | Viewed 1251 times ]


Figured it was time to revisit it to take care of a couple problems.

1. The low frequency response is not all that great even though I'm using push pull and the DC currents should cancel themselves out.
2. The output is limited to 50Vrms loaded with a 2.4K load due to that being the maximum voltage the transformer can handle.

A few things about the amp.

1. B+ can be anywhere from 20-40Vdc provided the 10K resistor is adjusted to keep the idle B+ current to 75mA.
2. The higher I go in B+ the higher maximum output I can get.
3. The no load output voltage can be much higher than the 50Vrms loaded voltage.
4. The diodes are resting against the chassis between the two power transistors with some heatsink grease to ensure they are coupled properly to the chassis far as heat is concerned as that keeps the amp gain more stable as the transistors heat up given the diodes are somewhat thermally sensitive.
5. The caps are used to eliminate fedback and seemed to work their best where they are conencted. Also all signal wiring is shielded which also helped to eliminate feedback.

Now to the issues at hand.

1. How would I increase the bass response as there's a larger high impedance 1920's speaker (uses cloth as the cone I think) that does have some decent bass response to it?

2. Would it be worth it to have Edcor design an output transformer that can handle at least 100Vrms with the same 600 ohm to 2.4K impedance?

3. How important is the DC resistance of the transformer windings that have DC passing through them?

Looking at the circuit the top darlington transistor draws 36mA and the bottom darlington transistor draws 39mA. Could that be enough of a current imbalance to cause the transformer to be lacking in the low frequency response or is it because I'm using push pull transformers that were never designed to have DC current through any winding?

My original thought was to find one of those 1920's speakers with better bass response then build a second amplifier and make a display for the four 1920's speakers I have with switches to select which ones conenct to the amp with the larger speaker serving to reproduce the bass down to 40-50Hz and the rest handling from somewhere around 200Hz on up. Would require a crossover though and I could build one that would work Said display would incluse an AM tuner (perhaps a good use for that MK-484 circuit) and an external input and maybe a bluetooth receiver.

That sort of display would be on down the road a bit as I really have no space to set up such a display which I would want large enough to accomodate at least 20 1920's speakers.

I do want the amp to be completely done though before I even think of such a display.


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Nov Mon 19, 2018 11:07 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
So I decided to play around with the amp today and took some measurements. I then then thought wonder what would happen if I added negative feedback.

So I experimented with it and came up with connecting the collector of one transistor to the base of the other transistor through a 100K resistor or a 50K resistor and did the same with the other transistor.

I took measurements before I realized that the input voltage was changing with frequency so I have no actual measurements at this time, but will take some tomorrow.

The basics are this.

The frequency response improved quite a bit in the treble. Bass response with the 100K resistors is good to 40Hz before the sinewave starts to be altered and with the 50K resistors is good to 30Hz before the sinewave starts to be altered.

The voltage difference between the output with the 5K load and no load is much less.

I did have to readjust bias as I used no coupling caps in the feedback circuit, although I could easily add some. Might experiment with that tomorrow.

I tried taking the feedback from the secondary of the output transformer, but connecting any terminal of the output transformer to ground only caused feedback.

The amp itself is inherently unstable due to the high gain I'm getting from one transistor stage and all signal wiring had to be shielded along with a couple caps added to keep the amp from breaking into oscillation.

I think that if I were to build the amp on a chassis with more room so everything isn't so compact the amp wouldn't be as prone to unwanted oscillations.

As I recall I even had to wire one transformer the opposite way to keep the amp from breaking into oscillation due to inductive coupling even though the transformers are around 3-4" away from each other.

I really wanted to include the transformers in the feedback loop as that would be better, but I don't see a way I can do it without the amp breaking into oscillation.

I may try to ground one end of the output transformer secondary and connect the other end through a resistor to the center tap on the input transformer primary and see if that works.

If I get oscillation I can easily unground the primary of the input transformer and try it like that.

Here's the schematic showing the feedback resistors. The resistor values are not marked as I am not sure which value I will go with yet. While testing with the feedback I'll try removing the two caps and see if the amp is still stable.

Attachment:
Amp 13.png
Amp 13.png [ 33.49 KiB | Viewed 1108 times ]


Now I don't care about bass response given this will be driving 1920's speakers which have no real bass response except for that one larger speaker that I think used cloth as a cone or something similar.

Would love to find one of those.

I would then build a second amp and find a cabinet so that I can display my 1920's speakers and have them working with the larger speaker providing the bass to 40Hz.

The idea was to use a MK-484 one chip AM radio as the source with a simple two way crossover then each speaker would have a switch that would turn it on or off.

See I like the 1920's speakers and collecting them, but I see no point unless I have a good way to use them and get good sound from them. I've never been one to buy things unless I can use them.

Even if nothing becomes of this amp I will at least have the experience of what I learned from building and experimenting with this amp.

I should name it The Muntz as I'm trying to get proper operation with the least amount of components. :lol:



EDIT:

Here's the data

Attachment:
Data.png
Data.png [ 17.83 KiB | Viewed 1071 times ]


Next I'll try the feedback from the output to the center tap of the input transformer primary.

For some reason when I used a different scope that provides more data I got no oscillation when I grounded the secondary of the output transformer.


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Nov Wed 21, 2018 4:39 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
I noticed something.

Across the 10 ohm resistor when I feed a 400Hz audio signal to the amp there's an 800Hz sinewave (happens at all frequencies) that is about 40-60mVpp which has a normal cycle then a slightly larger in amplitude cycle.

I can place a 470uF cap across the 10 ohm resistor and the waveform is reduced with the output jumping up by around a volt.

I'm assuming that waveform is due to the slight imbalance of the voltages on the transistor due to the DC resistance of the output transformer, but since the signals going to the transistor bases ate 180 degrees out of phase with each other shouldn't I be seeing just a 400Hz waveform?

Now to reduce that wabeform at 100Hz or below it could possibly take a cap that is over 2200uF.

Since the L shaoed chassis has screw holes on the small part that is bent 90 degrees I am seriously thinking about getting a piece of metal then removing all but the transistors and bias circuitry then moving them to the larger piece of metal which will allow for me to move things farther apart to reduce the chance of unwanted oscillation.

EDIT:

Due to the added feedback I was able to eliminate the two caps on the output transformer primary.

I experimented with a 470uF cap across the 10ohm resistor and that seemed to improve the frequency response somewhat and improved the linearity of the waveform near the transformer's 50Vrms rating.

Without the cap as the frequency went up the signal across the 10 ohm ressitor increased in amplitude.

I'll find a few different values of cap and see which value causes the signal to stay the same level down to 40Hz.

Here's the updated schematic.

Attachment:
Amp 14.png
Amp 14.png [ 32.37 KiB | Viewed 1050 times ]


The top transistor draws 36mA and the bottom transistor draws 39mA according to the voltages taken when I had B+ at 24Vdc.

Will need to take some new measurements with a B+ of 30Vdc which is what I use now.


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Nov Mon 26, 2018 2:58 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
I tried a 5600uF cap and it seems to work well.

I found a sloped front metal project case I had bought a few months ago which was intended for something else. I might install the amp in it. If I do that I will add a 30Vdc regulator for the B+ and a power supply if I have room.

That way I can power it from an external AC transformer.

Tried to take voltage readings at 30Vdc B+ but the bais wouldn't remain at exactly 75mA as it would wander around 74.8-75.5mA slowly.

One problem is how I have the bias compensation done with the three diodes touching the metal chassis between the two transistors. That is a compromise as any bias change is delayed by the time it takes for the heat to travel to where the diodes are.

Here's what I'm thinking. Move the diodes to where one each is directly touching a transistor case then leave the third one where it is.

Only problem is it is working off the voltage change with temperature of all three diodes so will the two diodes changing voltage more quickly than the third really affect it? In a way I'm only guessing about the diodes being on the transistors based on what I saw in one amp where a diode was right on top of a driver transistor.


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Nov Mon 18, 2019 4:44 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
Just thought about this amp tonight.

Should I bother making the transistor currents equal or would there be no real benefit?

My thinking is by balancing the currents it would equal a net DC current in the output transformer of zero which would help the low frequency response some and may increase the maximum output before distortion.

I noticed the base voltage is slightly different as well by 4mV.

Now I think the easiest way to adjust the balance is by adjusting the value of the 1 ohm resistor which won't be easy given the low value.

That said should I try to adjust to the higher collector voltage or the lower collector voltage?


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Jan Wed 08, 2020 12:05 pm 
New Member
User avatar

Joined: Jan Fri 03, 2020 7:54 pm
Posts: 22
Location: Malaysia Penang
Hi,
I think ,assume both primary each section have same number of turn , The DC for cancel out need to be equal to prevent core saturation. Even both side of primary may have different resistance.

For the bass response i think , it also related to the primary inductance , versus impedance. For example a 300V : 30V power transformer work at 60Hz versus 30V :3V at 60Hz although they have same ratio, the 30V :3V transformer cant work with 300V unless to increase the frequency up to 600Hz to prevent core saturation or maintenance a same flux density on the transformer core.. So in another view if using 30V:3V transformer for 300V:30V application , it will work well above 600Hz which can be the reason for poor low frequency response.


Top
 Profile  
 
 Post subject: Re: Solid state 1920's speaker amplifier revisited
PostPosted: Jan Wed 08, 2020 1:10 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20692
Location: Warner Robins, GA
The resistances are slightly different I believe.

That said the transformers used weren't designed for how I'm using them so not as much attention may have been paid to both halves of the transformers being the exact same resistance.

Not that important to have the bass response good to 20Hz as most 1920's speakers don't go nowhere that low in their frequency response.

Just wanted to be sure it was done right.


Top
 Profile  
 
Post New Topic Post Reply  [ 7 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 2 guests



Search for:
Jump to:  






























-->


Privacy Policy :: Powered by phpBB