Help on assembling an Aikido 5687 kit

[Nick]

So we're agreed that Broskie is using a very high bias voltage, too near 0 in my opiunion when my DS is emitting 2.3V.

But remember there will be a potentiometer in front of this, so you won't be seeing the full DS voltage at this point.

[simon]

...shows that the load diagrams are possibly not as accurate as simply using Ohm's law.

Yes, but don't forget that valves aren't completely linear - look how the grid volts lines curve. Perhaps more importantly you also need to consider how representative those curves are of your valves; some valve types are more consistant than others, but all valves age anyway and their characteristics will change anyway. The beauty of valves is that in many cases near enough really is good enough.

Broskie used a different idle current and lower resistor values for the cathode resistors in the upper valve.

No, he doesn't, look at the values for R3 and R4, and R6 and R7. The current through the first valve, V1, must be the same for both triodes, the ones with R3 and R4 for cathode resistors; similarly for the second valve, V3. So for an HT of 150V he has 7mA passing through V1 (if the current was different then some of it would be "escaping" somewhere). As both triodes are biased the same, R3 and R4 are the same.

He's biasing the second valve, V3, at 10mA so the cathode resistors are different to V1, but the same for both triodes.

You suggest that you would use the same resistors in the upper and lower valve.

Yep, you have no choice, otherwise it would be a right old mess!

It does present a small problem, namely Broskie supplied 4 resistors of each value, and with 2 for the two resistors on the lower valve on channel A and two for the lower valve on channel B, there are none left of the same values for the upper valves at all. I could simply order new resistors. But if you people pulled 15 mA out of the air as a reasonable current for the lower valve, could you possibly pull a different higher value out of the air which is just as good and I can use it for the upper valves? Then I can calculate which of his resistors best match.

If you think you might build more stuff, especially if you're not going to build something directly from a schematic then everyone here would strongly recommend you buy a resistor pack, something like this:
http://www.rapidonline.com/Electronic-C ... -kit-65195
But if you want to use the resistors you already have then try 14mA, 16mA, 13mA, 17mA etc to see if any you have would work.

Secondly, if we are now prepared to ignore Broskie's recommendations, shouldn't I simply use a different current and bias voltage from what he recommended for 150V B+ raw and see whether I can get 15 mA out of 75 V per triode with a decent bias voltage?

It's not quite that simple though sadly. Again if you look at the curves, at 75V you're limited as to how much variation in bias voltage you can use. The curves are an idealisation of the valve's physical properties - you can't go against this or you'd effectively be using a different valve.

[machtman]

Dear Simon
Thank you very much for your detailed comments.

I think we're closer than you realize. I use valves to refer to a pair of triodes in the input valve and a second pair of triodes in the output valve. I use triodes to refer to the individal bits within each valve. So, yes I realize that the resistors for each pair of triodes, R3 + R4 and R6 and R7, are identical and need to be identical within each pair. What was bothering me was that Broskie used different idle currents for the output valve and the input valve. As you correctly stated he used 7 and 10 mA. I'm using 15 mA for the input valve both for 240 and 200 raw B+ and I've been trying to get an estimate from you of what idle current I should be using for R6 and R7 on the output valve. Until now, I thought you were saying that you used the same idle current for both valves but since you seem to be using the designation valve where I would be using the designation triode, possibly I misunderstood you.

To make sure we're talking the same language, please let me know which value of resistors you ended up using for R6 and R7 with 300 V raw B+? And please indicate what your calculated idle current was on the two output valves and four output triodes.

Thank you for your suggestion of where to buy a kit of resistors. I'll look into this once we've decided on what else has to be changed as part of comparing 300, 240 and 150 raw B+. Until now, we need to change R12a and R12b as well as R17 as well as needing to connect with and without a central tap. I need multiple resistors for the cathode resistors on all four valves and 8 triodes. I'm assuming there is still something else that needs to be calculated.

I will be trying to read Morgan Stanley in the next few days and will be measuring all the resistors I have. Presumably I can also measure the caps. Is this also true with a standard multimeter for diodes? Do I need to worry about polarity when measuring electrolytic caps?

And it really is starting to make a lot more sense as a result of this introduction.

Thanks again
Mark

Nick
thanks for reminding me about the pot. I still think -2.8 sounds suspicious but have decided to follow Broskie religiously in his recommendations for 150V raw. Otherwise I can't dis his recommendations if I change them. I will also follow Simon's recommendations religiously on 300V raw but he is possibly more sensitive than Broskie, and he does talk to me, so I won't be dissing him oops: . Although I might make gentle comments on how it sounds. Now if you were designing this yourself and didn't have to worry about Broskie and Simon and had decided on 240V raw B+, what idle current would you be using on the output valves given 15 mA on the input valves? I'll check your numbers but trust your combined experience much more than my simple implementations of Ohm's law. As we've just been reminded, load curves are not linear, which negates Ohm's Law immediately over the non-linear parts.

I'm travelling to Edinburgh tomorrow afternoon and will first get back late Wednesday. So I won't get too far before the end of the week but will take Morgan Stanley with me.

all my best to all active participants

Mark

[pre65]

but will take Morgan Stanley with me.

Mark

Morgan Jones might be better.

[pre65]

[quote="machtman"]As we've just been reminded, load curves are not linear, which negates Ohm's Law immediately over the non-linear parts.


I'm not sure what you are suggesting there.

[Nick]

As we've just been reminded, load curves are not linear, which negates Ohm's Law immediately over the non-linear parts.


I'm not sure what you are suggesting there.

I am guessing he means you can't expect twice the voltage across a valve to produce twice the current.

[machtman]

Morgan Jones instead of Morgan Stanley

you might get twice the voltage for twice the current if you were on the linear part of the curve but each of the loadiing curves is sigmoidal and the bottom is changing slope continuously so you can't predict from Ohm's law what you get.

Nick, could you possibly do the calculations you would use for R6 and R7 so I can start soldering when I get back from Scotland?

Simon, I hope I haven't been too aggressive. I really am very grateful for all your help.

Till Wednesday night
Mark

[Nick]

Morgan Jones instead of Morgan Stanley

you might get twice the voltage for twice the current if you were on the linear part of the curve but each of the loadiing curves is sigmoidal and the bottom is changing slope continuously so you can't predict from Ohm's law what you get.

Nick, could you possibly do the calculations you would use for R6 and R7 so I can start soldering when I get back from Scotland?

Simon, I hope I haven't been too aggressive. I really am very grateful for all your help.

Till Wednesday night
Mark

I will have to have a read of the docs to see what R6 and R7 are.

[simon]

Mark, why not have a go at trying to work out the values yourself, you seem to be following the calcs and we'll keep you right .

To make it a bit easier, if you want to use the resistors JRB has supplied the nearest "spares" are 300 and 470. As the HT is the same, one will give you a bit more current through the valve, and one will give you a bit less. As a hint try 18mA and see what you get (something like a grid bias of -5.5V, cathode resistor of 305R, close enough for jazz).

You need to make sure that you're not putting too much current through the valve, and you're not as the 5687 will take in the order of 25mA per triode at 150V (from the link I posted to the full valve datasheet, power per triode = 3.75W = IV, so current = 3.75 / 150).

You also need to consider the current rating of your mains transformer, which is 115mA. But if you use a bridge (rather than centre tapped) and cap input (which JRB's board is) then you need to derate the current by a factor of 1.61 (look back to the links I posted detailing mains transformers).

So the actual current capacity of your 369AX as cap input bridge rectified = 115 / 1.61 = 71mA.

The first valve in each channel draws 15mA which leaves 41mA total for the second valves i.e. 20.5mA per valve. Happy days.

[machtman]

Hi Simon
I'm back and have a clearer head than a couple of days ago. I've been playing with the values. I've also been reading Morgan Jones, which is making more sense than the last time I tried, except for transistors which are still beyond me. Since last posting, 1000 single ended pins and the kit of 1000 5% 0.5 W resistors that you recommended have arrived. So I can solder and desolder to the pins and have extra resistors and don't have to worry about spares.

I've tried the calculations for 150 V, 240 V and 300 V raw assuming that I can use a full wave centre tap to get 175 raw (-25V for 150V) or a full wave bridge 350 raw (-50 V for 300 and -110 for 240). I've calculated the 150V for 7 and 10 mA as indicated in Broskie's suggestions and the 240 and 300 V for 15 and 18 mA. My calculations give me:

150V
R3,R4: 390R
biasL -2.7V

R6, R7: 240R
biasL -2.4V

R12a: 1600R
R12b: 3000R
parallel resistance: 1043R
R17: 200R
calculated drop 24.5 V instead of 25 desired

240V
R3,R4: 300R
bias: -4.5

R6,R7: 200R
bias: -3.6

R12a: 3900
R12b: 10K
parallel resistance: 2806
R17: 270R (from the new kit)
calculated drop 110.4 V rather than 110 desired

3300V
R3,R4: 390R
bias: -5.85

R6,R7: 330R
bias: -5.94

R12a:2000R
R12b:3900R
parallel resistance: 1322
R17: 100
calculated drop 50.2 instead of 50

The calculations were performed with R12a and R12b at 17 or 33 mA and R17 at 35 or 66 mA.

So my total current is less than the 71 mA you said I could get from the TX. However, to avoid soldering R12a, R12b and R17 multiple times, I was planning to solder them only once for 300 V and use the variac at 250 Vac and then simply lower the Variac to give the desired 220 V and 150 V for the comparisons. So I would start with the 300 V first and then the 240 and then the 150.

The kit contained one pair of 1 M R22 resistors and a second optional package of R22 from 124-300R. According to the manual, it looks like I should not be using R22 at al but replacing it by a jumper. Similarly, I should not be using C23.

I would like to start soldering soon. However,Broskie recommended starting with the diodes but provides different soldering plans for different forms of heater voltage. I have forgotten which type of heater power supply I am using, whether a fullwave bridge, full wave centre tap or full wave voltage doubler, so would be grateful if you could specify this again. It seems clear to me that I am using a full wave bridge or centre tap for the B+ voltage depending on the voltage I need.

Thanks very much
Mark

[Mike H]

Thicko question (sorry)

If this is a kit, why are we working out what component values to use to get such-and-such different Voltages?

[machtman]

Thicko question (sorry)

If this is a kit, why are we working out what component values to use to get such-and-such different Voltages?

Glad to know that somebody else doesn't understand everything. I was starting to think I was the only one. Especially because when I assembled the nixie kits, it really was as simple as do this and do that.

The Aikaido kit is a kit for experts, who apparently find it easy. It is also a kit that provides the opportunities to configure it as a head amplifier for headphones or as a line stage, which have somewhat different configurations. And most importantly, although it provides the on board power rectification that justiies the designation of all-in-one kit, it leaves the user free to choose any pair of transformers, or single transformer that fulfils the specs. Finally, although it is sold as a 5687 kit, it will also work with any valve with compatible pinouts. But possibly most important, Broskie doesn't dictate how it would sound best, leaving it up to the expert DIYer to decide just what bias voltage to apply, just what idle current to use, and all of this depends on critical resistors on the cathode. This has been an extremely steep learning curve for me, much steeper than I had anticipated, especially because I haven't even been discussing the options of what rotaries to use for attenuation or input selection, what wiring and solder to use for assembling it, or the input and output signal jacks and power connectors and case.

Hope this helps and I suspect the real experts might have a different view of how simple or complicated this is. I also suspect that Simon and Nick are supporting this thread because they think it is a good general introduction for multiple beginners in how to plan a line stage.

Mark

[machtman]

You also need to consider the current rating of your mains transformer, which is 115mA. But if you use a bridge (rather than centre tapped) and cap input (which JRB's board is) then you need to derate the current by a factor of 1.61 (look back to the links I posted detailing mains transformers).

So the actual current capacity of your 369AX as cap input bridge rectified = 115 / 1.61 = 71mA.

The first valve in each channel draws 15mA which leaves 41mA total for the second valves i.e. 20.5mA per valve. Happy days.

Simon
I was getting confused again and went back to Broskie's manual, which confused me even more. He doesn't talk about full wave bridge or centre-tapped about the B+ but rather on p.14 of the manual he talks about Centre-Tapped and Non-center-tapped, with the former being DC and the latter being AC. According to the description of the 369AX I downloaded from the internet, it should have 115 mA but the one that was shipped to me has the specifications 250VCT 100 mA DC for red-red/yel-red. (Broskie writes AC current is 1.2x higher than DC so this would be 120ma AC with the central tap). This listing of 100 mA DC is on the TX as well as on the box. So it sounds to me as if I may be pushing it with 66 because according to your calculation I should only draw a maximum of 62 mA. Unless my calculations were for AC current, it sounds like I need to either drop the second valve to 15 mA as well and simply drive R6 and R7 with identical resistors to R3 and R4 or drop the idle currents slightly on all four valves and recalculate everything.

I must also say that I am totally confused by an apparent discrepancy between your recommendations and the pictures by Broskie on p. 14. He shows a centre-tapped configuration at the top which leaves out diodes D3 and D4 and resistors R15 and R16. That produces 240Vct@120mA. The bottom non-centre-tapped transformer has all four diodes connected, which matches with your statement that one needs four transistors, but only indicates 120Vct at 180 mA. Based on this, it sounds as if the 100 mA specified for the TX might only be true when the centre tap is used and that there shouldn't be any problems with excess current. But I thought we were using the non-centre-tapped version to provide 250V whereas the diagram shows half the voltage versus the CT version.

Broskie is also confusing me with jumper J13, which always looks as if it is not connected to the chassis ground and I don't understand the meaning of the big arrow pointing down but not connected to anything.

Looking forward to this final clarifications and then I can finally start soldering (I hope).

Hope this isn't too many questions at once, but even though there are only 10 diodes to solder, I think I need to wait before I start until I understand what is going on here.

Mark

[Mike H]

Ah OK you reminded me of the options shown in the blurb, as regards end usage.

I agree it's not an ideal situation for a novice

[simon]

Right Mark, been rather bogged down for a few days - you've probably already got this thing built

I've calculated the 150V for 7 and 10 mA as indicated in Broskie's suggestions and the 240 and 300 V for 15 and 18 mA. My calculations give me:

150V
R3,R4: 390R
biasL -2.7V

R6, R7: 240R
biasL -2.4V

R12a: 1600R
R12b: 3000R
parallel resistance: 1043R
R17: 200R
calculated drop 24.5 V instead of 25 desired

240V
R3,R4: 300R
bias: -4.5

R6,R7: 200R
bias: -3.6

R12a: 3900
R12b: 10K
parallel resistance: 2806
R17: 270R (from the new kit)
calculated drop 110.4 V rather than 110 desired

3300V
R3,R4: 390R
bias: -5.85

R6,R7: 330R
bias: -5.94

R12a:2000R
R12b:3900R
parallel resistance: 1322
R17: 100
calculated drop 50.2 instead of 50

The calculations were performed with R12a and R12b at 17 or 33 mA and R17 at 35 or 66 mA.

Your figures look close enough to me, well done.The only one I might query is your bias for R6 and R7 for 300V - -5.5 looks a little closer perhaps? That would give Rk=305R. The good news is that valves are pretty tolerant, the Aikido especially, and within 20% for most things will be fine.

So you can build this with your choice of components, sit back and enjoy .

Now your Brucey bonus is your power supply. The figures for R17, R12a and R12b are perfectly valid, and numerically correct. But, the resistors and capacitors form filters that roll off the bass response. I don't know what values JRB supplies for C5, C7 and C8, but for the sake of this exercise let's assume they're all 100uF.

Simply, the bass roll off is given by the formula f = 1 / (2 x pi x R x C) where f is the frequency where the bass has fallen to half the full level. I've not explained that very well, this might help http://en.wikipedia.org/wiki/High-pass_filter

So you can see that in your first example with C = 100uF and R = 1043R,
f = 1 / (2 x pi x (100 x 10^-6) x 1043) = 1.5Hz

This is good - you might get different opinions but 1Hz for a pre stage is a good target.

But, R17 is 200R and this would give 8Hz.

So assuming all the values of caps are the same (I don't know if they are, but guessed at this earlier in the thread) you can see why keeping the values of R12 and R17 similar might have some advantages. Now whether you can actually hear a difference is a different matter entirely.

Don't forget to check the power rating of the resistors BTW, you can get a bit of a surprise if you don't. And for paralleled resistors of different values don't forget the current flowing through each will be different...

Incidentally what's the voltage rating of caps C5, C7 and C8? You need to make sure that they don't go over-voltage on power up before the valves have warmed up and start conducting.

The kit contained one pair of 1 M R22 resistors and a second optional package of R22 from 124-300R. According to the manual, it looks like I should not be using R22 at al but replacing it by a jumper. Similarly, I should not be using C23.

Correct and correct. Make sure you leave C23 untouched - don't add a link as that will just bypass the cathode resistor. R22 optimises the amp for use as a headphone amp. Nothing bad will happen if you do add one, it just won't be optimal.

Got to dash now, but will try to come back as soon as I can.