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


It is currently Aug Tue 04, 2020 9:15 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 9 posts ] 
Author Message
 Post subject: Phase splitter idea
PostPosted: Jan Mon 06, 2020 10:49 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20691
Location: Warner Robins, GA
While fooling around with an OP-AMP circuit to step up an audio signal to a higher level while being able to run on a 6.3Vac heater winding I came up with the following circuit.

Attachment:
OP-AMP 3.png
OP-AMP 3.png [ 14.73 KiB | Viewed 1033 times ]


So I thought and realized that I could drive a pair of output tubes directly with the circuit by removing the transformer.

And we have the following.

Attachment:
Phase splitter.png
Phase splitter.png [ 11.94 KiB | Viewed 1033 times ]


That could be useful for tubes such as the 6BQ5 and other tubes requiring a lower audio drive voltage.

The transformer could be used in instances where a higher drive voltage is needed, although one would want a transformer that is a step up such as 1:2 voltage ratio.

EDIT:

So it works fine, but

With the volume control full CW and the audio generator capacitor coupled it has a flat response to 20KHz.

24.5Vpp 400Hz
24.5Vpp 20KHz

If I turn the volume control a little CCW it starts to reduce the treble.

20.2Vpp 400Hz
11.3Vpp 20KHz

If I put in a 100K resistor to simulate a high impedance source it is worse.

22.2 Vpp 400Hz
9.20 Vpp 20KHz

If I set the gain to 1 I get the following.

Volume control full CW

23.2Vpp 400Hz
23.2Vpp 20KHz

Volume control a little CCW

18.1Vpp 400Hz
9.5Vpp 20KHz

100K series resistor

20.8Vpp 400Hz
8.8Vpp

I am not sure why I am having this issue.

I need to try a 10K volume control and see if that is any different.


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 9:52 am 
New Member
User avatar

Joined: Jan Fri 03, 2020 7:54 pm
Posts: 22
Location: Malaysia Penang
Hi,
Is nice circuit, i have make a similar op-amp circuit many years ago to drive 6J6(6N15) into push pull. I was using LM358,but i lost the op-amp circuit i cant remember. I think it will be nice to make use of 6.3VAC . I have make a second version by using 2SC945 as phase splitter , The reason is 6J6 is very small tube , i want the driver look smaller then power tube, So the only option will be BJT which is cheap and easy available. So i design a BJT as phase splitter by using 6.3VAC to multiply to -12V and the positive voltage is get from B+ regulate by zener .

Attachment:
6j6_pp-1.png
6j6_pp-1.png [ 23.34 KiB | Viewed 949 times ]

https://youtu.be/3k8jxa0s788

i have record my experiment into my page.
https://geeseang.wordpress.com/mini-hyb ... amplifier/


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 10:24 am 
New Member
User avatar

Joined: Jan Fri 03, 2020 7:54 pm
Posts: 22
Location: Malaysia Penang
Hi,
I have compare the LM6142 Bipolar op-amp and LM358 , LM6142 having higher input bias current, maybe due to the enhance slew rate circuit ,it get improve when LM6142 supply voltage raise up . I think both have same pin out , maybe you can try replace LM358 ? or TL082 which is Jfet op-amp , but TL082 and TL072 they cant work at single power supply. your circuit have dual supply i think should be ok.. The problem for LM358 is having lower product bandwidth. it may no have sufficient gain for high gain high bandwidth.
Attachment:
Op-ampcompare.png
Op-ampcompare.png [ 61.74 KiB | Viewed 948 times ]


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 1:09 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20691
Location: Warner Robins, GA
So would the TL-082 given it is a JFET input resolve the issue of the reduced treble?

Now would it be possible to put a JFET before the existing OP-AMP and achieve the same results?

Perhaps that would provide a little bit of gain as well which reduces the amount of gain I need from the OP-AMP.

As long as the generator is connected directly to the OP-AMP input or through a capacitor, the OP-AMp has a flat frequency response to well over 100KHz when the circuit is set for unity gain.


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 1:28 pm 
New Member
User avatar

Joined: Jan Fri 03, 2020 7:54 pm
Posts: 22
Location: Malaysia Penang
I cant 100% sure if using TL082 it can solve the issue. i just try to give my suggestion , if you have TL082 or LM358 you can try to replace and see . It seen like the current op-amp have lower input impedance, I guess there also have a capacitive load on the input pin, which when you adjust your volume CCW it increase the source impedance and the source impedance and input load capacitor(inside op-amp) form a low pass filter.. I'm just guessing , you can try and see by changing opamp see it can solve the problem.


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 1:52 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20691
Location: Warner Robins, GA
I may try the JFET idea first as that will be easier than desoldering the OP-AMP.

Plus if I use a TL-082, I won't have quite as much maximum output as the OP-AMP I used is a rail to rail type in that its output can swing extremely close to the power supply voltage.

EDIT:

I did a test of the TL-082 OP-AMP and with unity gain I was having the same issue.

I then decided to test just the 1 meg volume control which in this instance is connected VIA a shielded cable to the OP-AMP and to the input.

Seems like either the control is reducing the upper audio frequencies or it's the cable.


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 7:58 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 909
Location: Burke, VA 22015
The way the potentiometer is connected at the non-inverting input of the first op-amp does a poor job at attenuating the input signal. When the potentiometer's wiper is at the grounded end, the input is shorted to ground via the 0.47uF capacitor. This is bad for a source with a very low output impedance and not capable of driving large capacitance load. When the wiper is slightly above ground - let's say 1K above ground, the input is loaded by a small resistance (1K) but attenuated very little since the op-amp's input impedance is still much higher than the 999K resistance between the wiper and the non-inverting input. However the op-amp input and stray capacitances - let's say 10pF - and the 999K resistor will form a lowpass filter with a cutoff frequency of about 16 kHz. With a high-impedance source or higher stray capacitance, the cutoff frequency will be lower. This likely explains what you saw. I would use a 100K potentiometer for volume control to bump up the cutoff frequency by a factor of 10 to avoid this issue.

In general, using a large feedback resistor without a compensation capacitor will lower the phase margin of the op-amp. Since the maximum gain of your amp is 6, I don't see a need to use 1M resistor in the feedback path. It would be better to use something smaller such as 50K for the feedback resistor and 10K for the resistor between the inverting input and ground. Below is a modified schematic which shows new resistor values and correct volume potentiometer connection.

This phase splitter works. However it's not balanced. A better one would use three op-amps where two form a cross-coupled differential output stage such as the topology used in the Analog Devices SSM2142 and TI DRV134/DRV135. Besides the ability to drive high-capacitance load, the later can drive up to 17Vrms into 600 Ohm.


Attachments:
Phase_Splitter_Mod.png
Phase_Splitter_Mod.png [ 7.72 KiB | Viewed 894 times ]
DVR134_DVR135_BlockDiagram.jpg
DVR134_DVR135_BlockDiagram.jpg [ 80.4 KiB | Viewed 894 times ]

_________________
Binh
Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Wed 08, 2020 8:26 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20691
Location: Warner Robins, GA
I somehow managed to goof up the pot wiring on the schematic and didn't even catch that.

Thanks for catching it.

The output stage of the device I was going to use this with using the transformer to make the output unbalanced is as follows

Attachment:
Circuit.png
Circuit.png [ 12.7 KiB | Viewed 874 times ]


How would the 100K resistor affect the de-emphasis circuit?

I may try changing the feedback circuit components first and seeing what that does.

I tested just the 1 meg volume control and if I turn it a little CCW, the output drops as I go up in frequency.

Putting a 100K resistor in series with the AF generator to simulate the output of the device shows the same issue.

Looks to me like changing the feedback components isn't going to be enough so what I may do is just use a 1 meg resistor in place of the pot then install a 25K pot on the transformer output.

A quick check on the scope shows the circuit I did as very nearly if not nearly perfectly balanced.

Also with the generator connected directly to the OP-AMP input, I can get itn up to a frequency of 214KHz with no drop in signal, although any higher than that starts to distort the signal somewhat.

For testing the circuit I'll place a 10K resistor for the feedback and just set the 1 meg resistor very low as it would be easier that way versus having to desolder it from the perfboard to replace with the right pot.

EDIT:

Switching to the values in the feedback circuit you suggested definitely fixed one issue with the upper audio frequencies being reduced.

I then tried the 1 meg volume control and still had the issue.

I also tried just the shielded cable tying both wires together and using a 1 meg resistor between the two wires and the ground wire of the cable and still had the same issue.

Also tried removing the cable and just having the shielded cable on the input. Same issue.

I tried the cable to the 1 meg pot without the input cable and same issue.

So it looks like that for my particular application I would need to use the volume control on the output of the circuit or figure some other way or circuit.

Here's the modified circuit.

Attachment:
Phase splitter.png
Phase splitter.png [ 12.2 KiB | Viewed 840 times ]


I cannot understand how cable capacitance could be an issue even with only a 100K volume control.

I've used a 100K volume control in other circuits even tube circuits and I never remember cable capacitance being an issue at all. Plus I've seen higher resistance controls used in tube amps and receivers and I never noticed any cable capacitance issues.

EDIT:

So I found a solution.

I already have a 25K volume control on order so I tried a 20K trimpot I had in place of the 1 meg pot. That made the frequency response flat to a little above 20KHz even with the shielded cable.

25K is no good for vintage sources which have a higher output impedance, but would be just fine for any modern source.

A solution for using this with older higher impedance sources is to use an OP-AMP set for unity gain which then feeds the 25K volume control. I tested that and with a 1 meg resistor from the + input to ground I had the same flat frequency response.

That would ideally require a quad OP-AMP which would give an extra OP-AMP.

If one were concerned about the outputs being balanced perfectly the circuit could be modified to where the first OP-AMP is a buffer, the second one provides the gain and the third and forth are unity gain with one inverting the signal.

Or perhaps the better circuit posted here could be implemented with the quad OP-AMP and the one OP-AMP being the buffer.

One thing I may experiment with is the buffer to see if any gain can be added to it without affecting frequency response as the circuit doesn't quite have enough gain to put out its full undistorted output with a lower output like some MP3 players and cellphones have.

I can get full output at 1Vrms input though. Adjusting the OP-AMP gain for full undistorted output still gives a flat frequency response that is 29.8Vpp at 20KHz and 29.2Vpp at 147KHz with the 20K pot adjusted to provide 29.8Vpp output with the max undistorted output being 34Vpp.

Setting the OP-AMP for max gain and adjusted to provide 30Vpp output at 400Hz I get 30Vpp at 20KHz and 29Vpp at 124KHz.

Now with the buffer added and a 100K resistor in series with the generator to simulate a high impedance source I get the following when the OP-AMP is set for max gain. At 400Hz I adjusted the 20K pot for 30Vpp. 20KHz I have 30Vpp. At 35KHz I have 29Vpp.

Attachment:
Phase splitter 2.png
Phase splitter 2.png [ 14.57 KiB | Viewed 831 times ]


bb.odin wrote:
A better one would use three op-amps where two form a cross-coupled differential output stage such as the topology used in the Analog Devices SSM2142 and TI DRV134/DRV135. Besides the ability to drive high-capacitance load, the later can drive up to 17Vrms into 600 Ohm.


http://www.ti.com/lit/ds/symlink/drv134.pdf

Looking at the datasheet I like that chip. I may order one from Mouser as I need to order a panel mount F jack and a 75 to 300 ohm balun.

That has an input impedance of 10K so the OP-AMP buffer would be needed to provide some gain provided the gain doesn't cause any reduction in frequency response.

Because I already have the circuit built I will experiment further with the circuit to see if I can get it right.

I may later on work with that chip and design something that is better.

At 17Vrms output that would give 48.0828Vpp which can drive most audio output tubes including the 6L6 in class AB1.

That requires a +/- 18vdc supply though and for working with a 6.3Vac winding that wouldn't be exactly possible, although Traco may make a +/-18Vdc switching supply that may work on the rectified 6.3Vac winding.

I normally don't combine solid state with tubes, but something like this makes for an easy two tube amp that is very compact.

I received the transformer today and will try it with the circuit tomorrow.

I suppose that I could use a step up transformer to make the output high enough to drive more power tubes.

With that driver chip if I used a 1:2 transformer I would have 96.1656Vpp which can drive the 6L6 in class AB2.

So that chip can make a very powerful two tube amp.


Top
 Profile  
 
 Post subject: Re: Phase splitter idea
PostPosted: Jan Fri 10, 2020 2:01 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 20691
Location: Warner Robins, GA
Here's the circuit as I'm currently testing it.

Attachment:
Phase splitter 3.png
Phase splitter 3.png [ 17.32 KiB | Viewed 818 times ]


Seems to work quite well, although there is a phase shift at 20KHz of 50 degrees with the output leading the input and the phase shift is 1 degree at 400Hz.

Now purists may complain about that, but in my experience using those transformers I haven't noticed any audible effects of the phase shift.

I did notice that it seems like the high frequency performance is actually better.

What I was doing before to get the total peak to peak output was summing the two scope channels and perhaps that wasn't quite right.

So because the transformer has a 15K secondary I used a 15K resistor as the load.

The phase shift is still the same with no load, although the output jumps up by 2.66Vpp when the load is removed.

I am thinking of trying that all in one phase splitter chip. There is also a receiver chip that converts the balanced signal to unbalanced.

That chip would make it very easy to convert an unbalanced source to balanced.

EDIT:

Tried the circuit tonight with the intended amplifier (has its own phase splitter) with the transformer wired to do a balanced to unbalanced conversion and I fed a 400Hz signal to my FM RF generator and set it to the correct level so that the output of the tuner would be the same as it is with a strong FM station then set the volume control until the amp just did go into distortion. I adjusted the 50K gain pot until the amp would just go into distortion at slightly less than full CW on the volume control.

Seems to work good.

Now because this is a phase splitter I decided to try it with a pair of output tubes.

I have a 10 watt push pull 6V6 amp that I use with my Hammarlund SP-600 JX-17 receiver with the receiver's output transformer providing the phase splitting and being able to drive the amp to full power easily.

So I decided to test the amp with the circuit to see if it indeed could drive the amp to full power and it just did come shy of the necessary signal level, but if I were to use an Edcor XSM series transformer of 600 ohms to 2.4K which is a 1:2 ratio and that all in one phase splitter chip, I could easily drive that amp to full output from a circuit that receives its voltage from a 6.3Vac heater winding.

That would require another OP-AMP though so a normal line input could be stepped up enough to drive the chip properly.

I could see two of those chips being placed in a box with a +/- 18 Vdc supply and an OP-AMP to provide initial gain along with a stereo volume control and a double pole multi position switch for several inputs which would be the preamp. One could then build a pair of monoblocks only using two output tubes each and have a very nice stereo amp. Given the signal to the amp is balanced, each amp could be located at each speaker.

Now if a transformer with a voltage ratio of 1:3 is used, it is possible to get an output of 50Vrms (max the XSM series transformers are rated for) which is 141.42Vpp and that can drive most any power pentode to full output and even some power triodes. The closest to 1:3 is a 1:4.1 ratio which is 600 ohms to 10K.

To eliminate any chance of hum pickup by the transformers, they would be located in the preamp box.

So in effect I have just created a preamp that can put out 50Vrms into a 10K load.

Now if I were to have a transformer made capable of handling a higher voltage I could get a maximum of 69.7Vrms which would be 197Vpp into a 10K load.


Top
 Profile  
 
Post New Topic Post Reply  [ 9 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: Ed Kraushar and 14 guests



Search for:
Jump to:  


































-->


Privacy Policy :: Powered by phpBB