Magnavox Flea Power: Getting More Out Of The 8600 Series - A Lot More!

[Wharfcreek]

JD, I'd also mention that in Dave's comments somewhere WAY back in this thread, he discusses the fact that a tube shield over the driver tube is also advisable. I think he shows a picture of the one he used where it had a small pigtail wire to establish a good ground on the shield by connecting it between the shield and soldering it to the chassis. When I did my 'scratch builds' I contacted Dave about this as I wanted to avoid the shield. His advice was to make sure the driver tube was at least 2" away from each output tube. In looking at your build, your driver tube looks rather close to the output tubes, and as such, I'm wondering if the addition of a shield might be beneficial to your amp's operation? Again, I think this is a 'stability' issue.....but I think Dave indicated it had some HF stability function.....so while it may not be audible, it may fall into that 'measurable' spectrum of things. As to your buzz, is it there if you remove the pre-amp leads and use those 'grounded' cable/plugs?

 

[gadget73]

.and then using a 100 ohm pot vs the 50. This would have provided a bit more 'range' on the bias adjustment

The problem with having too much range on the adustment is that it gets touchy. If you need a wide range to get sketchy tubes to work right, you probably just need tubes that aren't so sketchy. New EL84's aren't all that expensive to buy a pair of, and even old ones can be had by the pair for not insane prices.


As for the proximity thing, you can actually measure cross-channel interference from drivers too close to an output tube. Dave was helping me with my Pilot amp, and my measurements confirmed his testing that the driver tube common to both channels would actually pick up signal from the nearby opposite channel output tube and bleed it through the amplifier. Don't know if you'd really hear it, but its certainly easy enough to measure and fix.

 

[Wharfcreek]

Hence the shield?

 

[gadget73]

Yep. The shield made the crosstalk stop. Not positive if thats the reason for it in the Magnavox, but I suspect it is. Easy enough to determine either way, drive one channel to full power and put a scope on the other channel. If you get signal with no input that goes away with a shield, you know the answer. With the Pilot it wasn't a tremendous amount of signal, but it was a non-zero amount that went down significantly with a shield.

 

[jdsalinger]

I'm confused about grounding the common on the OPT. I didn't have to do that on my KT88SE build. Why this one?

The problem is that, today I was looking at the performance using my MPI-4. I wanted to see if there was any crosstalk. I was shocked to discover that there is NO stereo separation due to the green wires from the OPT being tied to chassis ground (no continuity between inputs). I even confirmed that by pulling out one of the input interconnects and still saw a sine wave on the scope. The amp is in MONO!
Yet, tying the neutral taps to ground DOES tame down the gain, so it has to be done. No way around it.

I guess I didn't notice that it was mono when listening to it last night. Help!! What am I missing?

Next question, why is there a ground symbol on the laminations of the OPT and PT? I've never seen that before.

 

[gadget73]

Most all amplifiers have a grounded common on the output transformer, but its only absolutely necessary when there is a negative feedback circuit involved that taps off the output trafo secondary. With no feedback, strictly speaking you don't have to do anything but connect the OPT to the speaker. If there is feedback, not connecting the output transformer to ground will cause the feedback loop to not actually work, which you discovered already.

It just means that the laminations are connected to circuit ground. Usually its not spelled out since it gets grounded by bolting it down. You don't have to connect an extra wire or anything. Its probably just more for completeness reasons, or maybe in case someone wanted to build one on a non-conductive chassis that they know to ground the lam stack.

Connecting the common side only of the output transformers to ground will not cause it to run in mono. Both positive sides go to their respective speaker connections only. If yours is running in mono, there is another problem. Check to make sure the black wires aren't shorted together somehow, and make sure that the input jacks are grounded to circuit ground and not just to each other.

 

[jdsalinger]

Check to make sure the black wires aren't shorted together somehow, and make sure that the input jacks are grounded to circuit ground and not just to each other.

OK thanks.

I checked the black wires and measured 1.6 ohms ONLY when the greens are tied to chassis ground. Basically shorted, right?
This makes sense. Black and green are opposite ends of the secondary and if greens are both tied to a common ground, then resistance between black taps should be 2X the secondary resistance.
The original 86xx schematic shows both OPT commons tied to ground, so I guess this is correct.


It's possible that I'm making some incorrect assumptions here, and maybe I'm expecting something that isn't possible with the given schematic.

I applied a signal to only the R or L channel and on the dual trace mode, I could clearly see that either R or L trace showed a wave while the opposite trace was flat.
OK good, so it doesn't seem to be in mono. If it WAS, I would see waves on both channels simultaneously.

BUT, the MPI-4 has a function that indicates stereo separation. When a stereo signal is played and the channels are separated, the display traces appear as a circular "ball of twine" centered on the display....if there is " less than optimal " stereo separation, the traces fall onto a diagonal line. It can vary from oval (OK) to a very distinct bar (worse). I was getting the distinct bar.

So, the question is; Is stereo separation to be expected? I would assume so, because of the talk about proximity of the driver to the power tubes.

Or is it not that important & not to be expected?

Edit: I just read Dave's post regarding stereo separation. I may have to use the shielded socket method, since I placed my tubes very close together.

 

[jdsalinger]

First off, Merry Christmas to the DG-SE1 club!

This obsession just doesn't take a holiday. (actually, I'm a very early riser, and wife is asleep, so this is JD's time)

I ran the DG-SE1 through the MPI-4 test procedure again this morning.

First test is balance and phase check. Looked really good.

Second test, stereo separation. This is a somewhat subjective test as I described above. Separation is not really as bad as I first thought. I ran a pink noise track and I was able to compare the preamp performance against the power amp performance. Both looked pretty good, but the preamp showed a nearly perfect pattern. Then, I played a music video of a live performance and on the power amp, got a more or less wide oval shape that would alternate between circular and diagonal patterns. The instruction manual for the MPI-4 describes a definite diagonal pattern as indicating "less than optimal" stereo separation. So, what I was seeing is definitely better than that.
I don't use a shielded socket on the preamp tube at the moment, but I have a shield that I dropped over the top (it touches the screw holding the socket to the chassis, so there is some amount of physical grounding going on). I'm pretty sure I could see a slight improvement with the shield. I think it would be worth fixing that soon. I didn't realize that proximity of driver to power tubes could have ill effects. Live and learn.

Final test is a frequency sweep from 20-20KHz. This just came from a YouTube track, so maybe this isn't valid, but it's all I have. 20-50, the wave was really strange looking. 50-70, I could clearly see that the wave had a dogleg in it. I take that to mean that the OPT (the new ones from AES) just can't handle the frequencies down that low (saturated?). At 70Hz, the dogleg disappeared, leaving a smooth sine wave that was smooth all the way up to 17K-17.5KHZ, where it started to get shaky and then flatten out. The time it takes for that track to go from 17-20KHZ is about 3 seconds, so hard to say exactly where it stops, but my ears cut out at about the 14-15K range, so 17K is good with me.

My Three Pi speakers roll off below 50Hz (maybe even higher) so I have a sub to handle the LF up to 100 Hz anyway. I wanted an inexpensive SE amp to handle the highs and this one fits the bill nicely.

 

[gadget73]

Correct, you will read DC resistance from black to black, but thats because the two common leads are grounded. Just because it has low DC resistance doesn't mean its shorted though. You're just reading the coil resistance. If injecting signal on one channel produces output on one side and a flat line on the other, you're not running in mono mode.

The dogleg could be saturation, it may also just be that its beyond full power output. When you hit saturation, the wave form suddenly gets very distorted below a specific frequency and/or above a certain power level. You can typically raise the frequency a couple hz and it will snap to a normal looking output. High power, low frequency is where you see it happen though.

The bleed from proximity is an EM thing. Tubes do produce an electromagnetic field, and they are influenced by EM fields as well. Put the two together and you get cross-talk. Its not a massive amount, but it is measurable

 

[dcgillespie]

JD -- Great that you're testing your build! I always encourage that effort.

Gadget gave you the straight skinny on grounding the output winding of the OPTs. Basically, the NFB voltage is being applied to the non-inverting input of the 12AX7/6EU7 stage (across the parallel value of 820 // 150K). If the Grn lead were not grounded, there would be no voltage applied across this input, as surely as if the input interconnect were not grounded, then there would be no voltage applied to the input of the amplifier across the 470K grid return resistor.

As for cross-talk between the channels, there are four factors that can influence this issue:

1. Physical layout -- where lack of appropriate space between components allows for unintended coupling between them.

2. Use of a common power supply -- where inadequate decoupling can allow for coupling between the channels.

3. Lead dress -- where the wiring of one channel can be picked up by that of the other channel.

4. Coupling between the sections of the 12AX7/6EU7 AF Amplifier tube, since each section serves one channel.

Items 1&3 are controlled by details of the construction, item 4 you have no control over (but in reality is insignificant), while item 2 is addressed through proper design of the circuit. In virtually all applications, issues of cross-talk with an otherwise proven design will come down to some aspect of items 1&3. Where item 1 typically was never originally a problem, it can certainly be made a problem if oversized boutique components are crammed into a space never intended for components of that size. The unintended consequences are rarely ever considered by those using such components. In the great majority of cases, problems of cross-talk usually boil down to item 3, or some aspect of the build not executed correctly.

I hope this helps!

Dave

 

[dcgillespie]

I also meant to mention that testing for frequency response is always conducted at low power levels -- typically 1 watt. But in flea power units like this, such testing should be conducted at no more than .5 watt, with .25 watt being a very appropriate testing level to use. Using such a level ensures that the test remains a test of the circuit's frequency response, and does not spill over to become a test of power bandwidth, which is almost entirely defined by the characteristics of the OPT used.

Dave

 

[jdsalinger]

Dave,
That was good advice. I retested the frequency response using the lowest power I could. When the gain was too low, the trace got wobbly, so I increased gain just to the point where the sine wave looked fairly stable. Then the dogleg was gone. The sine wave wasn't really smooth below 30-32 Hz, but it was still recognizable as a sine wave. It smoothed out at about 42 Hz.
The signal flattened out at about 16.9 or 17KHz.

I rewired the heater wires with solid core and bundled HT wires to clean things up a bit. Still have just a barely audible 60Hz hum, but it's less than my Pilot 402, so I'm pretty happy. I'll take another look at the grounds at some point.

Still a little cramped and "home made" looking, but better.

 

[Son-of-Vere]

Just fired up my DC-SE1 tonite. While I have some crispy socket noises, and some PS voltage issues to work out, I have a conundrum. The bias voltages are too high and I'm out of adjustment. I was around 270v going to the OPT's and minimum bias was 6.2 volts. Can I add more resistance in front of the bias pot to compensate? And are 1/2 watt components sufficient in the bias circuit? My math showed .25 watt dissipation in the bias circuit.

loaded OPT's @8 ohms, inputs shorted, FYI.

I ran the amp on a bucker which got that 270v down to 244v, only then could I set the bias to 5.85v.

See this thread for build info. http://www.audiokarma.org/forums/index.php?threads/motorola-hs-706c.743508/

 

[Wharfcreek]

SOV, I'll add some commentary here though it may get contradicted at some point. First, I think your B+ voltage at 270 going to the OPTs is OK, but you have to remember that you're not using the P-T31 output transformer that Dave originally designed around. As such, your OTs may offer some different DC resistance and thereby affect the actual voltage shown at the plate of the output tube. So, I'd consider working your PS around getting that plate voltage in line with the spec on Dave's schematic as a 'first step'. From that point, I would next take a look at the screen voltage....which I believe if you're at 265 at the plate, should be 255 at the screen. I believe it is this relationship that affects the ultimate 49ma current draw that is the target for for this amp's design. To some extent I think this was a function of the original 8600's PS capability and Dave's desire to get this amp in as 'true' a Class A operational status as possible. So, 265 at the plate, 255 at the screen, and then a tube that is in the 'normal' operating range for the EL84 tube lineup. One thing I noted originally was that 'some' EL84s either would not reach the 49ma (5.85 V) range, while others could not be brought down to that same point. I had discussed with Dave the concept of changing the 120 ohm bias resistor to 100 ohms, and then changing the 50 ohm bias pot to 100 ohms. This would add some range to the bias adjustment point. In the end, I never did this......but I did have a friend who did....and his report was that it worked fine!! And, BTW, I think I calculated the power in this circuit to be about the same 1/4watt.......so, I think you're OK there. I use a minimum of 1/2 watt to be safe....and as I see you're using the little mini-trim pots, if you're not using at least 1/2 watt versions of those, I'd suggest you change them as at 1/4 watt you'd be right on the ragged edge of potential burn-out.

Anyway...I'd try swapping some tubes around and see if the situation remains. I'd check both Plate AND SCREEN voltages and make sure they're in line. And, I'd also double check those little pots to make sure they're not already damaged (particularly if 1/4 watt units).

BTW, and reason you didn't mount the choke up on the 'other' side of the chassis......perhaps on the other end above your AC wiring? With all the room you have on that side, seems like a potentially viable place to locate it so as 'not' to have it hanging off the bottom. Just a thought.............

Let me end by saying I really like what you're doing in this project!! I've been wanting to try a 6SL7 tube.....so I'll be anxious to see how this works out for you!! Nice job underneath as well!!

Tom

 

[jdsalinger]

Got my 6P14P-EV yesterday. They aren't even broken in yet, but this amp is....FUN !

Listening to some modern soul/pop/R&B on my 90 dB/1w/1m Advent 5012's this morning.
Preamp, computer and Spotify volume all at 50%.
Using a 12AT7 as a driver and it was as loud as I dared without waking anyone in the house.

Of course, Advent bass is legendary, but the deep bass on Sampha's "Don't Touch My Hair" sounded like the heartbeat of some underground giant.
I'm very pleased with these P-T31s. With the right speaker....no need for a subwoofer.

I was nervous about using the 270DAX power transformer, since it is only rated at 104mA, but the temp only gets to about 48C after about an hour.
Might get a little hotter at times, but that seems safe to me.
I would still probably go with more headroom, like Tom did, if I was to do it again.

Now that I know how worthy this circuit is....I'm making plans to rebuild and fix the "oopsies" I had on this pilot build.
I think I'll get a Hammond chassis & put some separation between the power and pre tubes. Will do an AutoCAD layout and wiring diagram to simplify it and clean up some more.
I want to install bias pots & use star grounds per Dave's suggestion (I missed that post) maybe try a better or smaller coupling cap too.



See what you guys did???

 

[Wharfcreek]

JD, Great Job!! And, interesting comment about your use of the AT in the front end. I too have to confess that I spent some time swapping ATs and AUs into the driver tube position......and it 'seemed' to me that depending on what sources I used, the different tubes seemed to react either 'better' or 'worse'. I ran this by DSG and his comment was that the design of the amp really was set up to use the AX and that use of other tubes might result in some unwanted distortion due to improper voltages, or so I seem to recall. I eventually gave up on playing around with other than AX tubes and just stuck with the AX. I've yet to achieve the ability to actually 'measure' distortion levels.....and until I do, I have to trust in Dave's expertise! This said, there's always going to be the issue of 'what the numbers say' and 'what sounds good to you'. This is an 'argument' that is only 'won' by the individuals final decision. I know a lot of guys that build guitar amps......where the issue of 'distortion' is actually a desirable aspect of the amp's output. But, some of these amps have things 'compounded' by having either pre-amp or output stages (or both) driven WAY beyond what 'book specs' say the tubes are capable of. In some cases this results in tube life being greatly shortened. BUT....... from the perspective of the guitar player, the 'sound' of these amps is EXACTLY what they want. So, 'specs be damned' in that case....but, the 'cost' is a new set of tubes about every 50 hours of play time! Rather expensive IMHO, but if I was making the kind of income that some of these 'rock stars' make....then maybe a new set of tubes is just the cost of doing business.....hell, maybe even a tax write-off.

Anyway.......glad it's working well. I'm still waiting for the PT for my next version of this amp. It's an addiction!......lol TSD

 

[dcgillespie]

Use of a different driver tube won't change the power output capability of the design, but it could in fact cause a slight coloration to the sound due to reducing the open loop gain (OLG) of the design. Besides the operating point of the output tube being precisely located in the center of its load line, the other element that maximizes the performance of the design is the significant amount of stable NFB that it uses. While any amplifier design can cause sound coloration unless corrective measures are used, single ended designs seem especially susceptible to this. Using driver tubes with a lower Mu certainly won't hurt anything, and you may even like the effect produced. But maximum performance will always be achieved when the specified driver tube is used.

Dave

 

[jdsalinger]

Good to know and thanks for the explanation.

 

[solderjunkie]

Thought this might be a good thread to ask this question:
Were the original inputs for this 8601 series amps configured to accommodate phono? In other words is the stock preamp essentially a phono preamp?

 

[Wharfcreek]

I believe the answer to your question is 'no'. I believe that Magnavox used this amp in basically 2 different configurations; one being 'Phono Only', and the other being associated with a 'tuner'. The distinctions between these two units can be noted in a couple of different ways, with one being that the unit had 'controls' mounted on the chassis (the 'phono only' unit). In the other case, the amp had a 'molex' connector that supplied power to the tuner/control section. In either case, the 'amp' itself didn't have what is generally considered to be a 'phono' stage. The 'gain' of this amp is not sufficient for use with a magnetic phono cartridge. Like any 'power' amp, it is generally necessary to use a pre-amp of some kind, even with an input source like a CD player. There are 'passive' pre-amps that would act in a similar fashion to what the 'phono only' version of the 8600 amp would do in that they would control volume, tone, and balance. However, like a passive pre-amp, no additional gain is generated. Hence the need for an additional stage, generally referred to as a 'phono' stage, is needed. When a phono pre-amp is used with a passive pre-amp, there is still no gain generated in the pre-amp section, only the previously mentioned control functions. However, when a phono pre amp is used with a conventional pre-amp (or 'active' pre-amp).....then additional gain is added. Most modern-day pre-amps are 'active' unless specifically sold a 'passive'. And most include a phono stage, however, not ALL do! The original Magnavox phono used a high-output 'crystal' or 'ceramic' cartridge. This didn't need the same level of additional gain that a 'magnetic' phono cartridge required. So, with just the on-board passive controls, the unit could generate sufficient volume levels with the combination of the ceramic cartridge and the single voltage gain stage on the amp (1/2 of the 6EU7 for each channel). However, if a magnetic cartridge is used in your phono system, then a 'phono pre amp' will still be necessary, along with some kind of 'control' set up.....either a 'passive' pre-amp, or a conventional 'active' pre-amp. Hope this helps. WC