Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 12:00:12 PM

Login with username, password and session length

Recent Posts

 

DM K-2OX OUTPUT/FREQ PLOT

Started by benzer, September 28, 2007, 12:30:23 PM

Previous topic - Next topic

benzer

I Did draw the whole amp (with biggest gain and biggest boost for bass/treble/midrange) by spice (clean channel)
and i put input of 0.02 volts for (ac analysis)
(+15 -15)op amps feeding
and i got this curve,,, bUT ::::
it looks like there is not enough power in the bass and midrange  and too much distorion at high frequencies (over 6khz) (output more than 15v)...
this is a clean channel so there should be no distortion and if we preserve that the gain at midrange and bass would be very little wich leads to low power...... that looks weird,,any body can explain?
Good OL' BeN

teemuk

First of all, it doesn't look all that weird when you use a logarithmic decibel scale instead of linear voltage scale. Secondly, it still will not look right so I suggest you check your circuit. There is an error somewhere. Especially make sure that the potentiometers in the tone control circuit are indeed in a proper orientation; this makes a huge difference. I simulated the preamp and got a very "healthy" looking response in both clean and OD channels and with tone controls in various settings. If that is indeed plot of the clean channel you seem to have way too much gain and a way too enhanced treble.

benzer

well i did check my circuit ,it has some caps values errors but it didnt change that plot much,,
its actually the tone control is cutting much of the midrange and bass gain (fender type)
and when this curve is getting amplified the treble area is goin way higher than the mid and bass.
i did change some values of the circuit. some caps and resistors untill i got this plot,,,look how HOT it is.
i killed that very high gain in the high freqs area and i gave some more gain to the bass area. keepin the midrange weak to maintain the properties of the fender tonecontrol.
the input here is 0.02v and again the circuit is workin with its highest gain with maximum bass/treble/mid boosting
,,i feel like the amp is very sensitive,,right? i mean when the input gets higher than 0.02 v there gonna start distortion...what do u think on this?
also im plotting it in the linear axe cuz i wanna know where exactly its gonna exceed 15 volts,,so i can modify to get rid of the distortion.
Good OL' BeN

benzer

i tried to upload the spice file but i couldnt,, its not allowed
ill upload it as a picture but its not clear at all...
well anyone feel free to ask if any value is not clear
Good OL' BeN

teemuk

Ah, I see you simulated the output stage as well. However, do notice that a speaker load is not equal to a constant resistance and this will have a considerable effect to frequency response in circuits that employ current feedback.

I also believe that the factory schematic contains (yet another) error: C12 should be 10 uF instead of 1 uF. As you see from the graphs (which plot the response of the output stage only) the higher capacitance introduces a lower cutoff frequency and allows the speaker's resonance to be accentuated. With a constant 4-ohm load the response is dead flat - aside the hi-pass filtering.

As you see, the current feedback setup inherently introduces some low and high frequency boost; this is an effect, which should be considered when you tweak the tone control circuit.

In the plot the graphs are following:
constant_4r_1u:  Constant 4-ohm load, 1 uF coupling in feedback
constant_4r_10u: Constant 4-ohm load, 10 uF coupling in feedback
1u: Simulation of a nominally 4-ohm speaker, 1 uF coupling in feedback
10u: Simulation of a nominally 4-ohm speaker, 10 uF coupling in feedback

The frequency response naturally varies according to speaker's/cabinet's electromechanical properties. I used a Thiele-Small-based model of Eminence Legend 875 speaker.

benzer

well i did simulate the power stage and i got the same results ,
and yea that c12 was one of the stuff i did modify,here are the values i changed

C2=15P      >>>   100P
C3=33000P >>>   0.5u
R5=100K    >>>   200K
VR1=1M     >>>   330K +500K(POT)
C8=2200P   >>>   2000P
R10=3.3K   >>>   2.2K
VR4=10K    >>>   5K
R15=470    >>>   800

i plotted the whole amp with only changin c12 from 1u to 10u
,,well, surely still there huge gain,so i did change the other values and i got that other plot in blue ,,the green plot is the voltage(before the output cap) reference to ground.
if my supply voltage is 15 i dont know how on earth the spice gives me voltages bigger than that ,,,!!!!!! so i did the whole modding to ensure that there is no voltage in the circuit is more than the supply voltage at maximum gains.thats why i did plot the voltage before the output cap(cuz its the biggest voltage in the circuit).
i guess i kept the characteristics of the tone and the amp in general .
also if you calculate the cutoffs at the first and second stage after the modding you see that they are (16hz) and (33000hz)
which i think is way better than (47hz) and (200000hz) days before.
im still afraide about boosting the bass that much that speakers blow,, is that possible,, maybe they did attenuate freqs below 50hz intentionally...
still dont get it why they designed an amp with such freq respons,,

anyways, where do u get speakers spice models from?
i would like to try that on the modified amp.
C11=15P    >>>   100P
C12=1u      >>>   10u
Good OL' BeN

benzer

thats the response of the amp without any mods at all
Good OL' BeN

benzer

Good OL' BeN

benzer

here is the all values modified plot ,,i made sure no voltage in the whole circuit goes over 15 volts..still dunno how spice plot such voltages....
the green is the node before the output cap reference to ground
the blue is the voltage across the 4ohm resistor
Good OL' BeN

teemuk

The assumption about that capacitor value is just that: An assumption. For the record, my little Crate practice amp uses the same power chip and therefore has nearly similar output circuit, only it uses a 22 uF capacitor and DC couples the output. There is a slight amount of HF rolloff put to the current feedback path and little bit more gain as well. Anyway, direct comparsion is still meaningful.

...Now the "funny" part, I simulated it and it turns out that the 1000 uF DC coupling in the output of K20X actually boosts up the lower bass. Anyway, the response is still "wimpy" if only 1 uF cap is used in the feedback. The input signals of all amps are matched to even the levels for comparison.

I think the only way to know for sure is if someone opens up his amp and checks out the value fitted in. I don't have much trust into the accuracy of the schematics.

The Thiele-Small-based model is explained here:
http://www.epanorama.net/documents/audio/speaker_impedance.html

For example, the Eminence Legend 875 model uses:
Re: 7.49
Le: 0.56m
Cmes: 265.63ยต
Lces: 10.24m
Res: 110.34

benzer

so you think my mods should worth building?
or its better to go back to the original circuit?

im really confused. :duh
Good OL' BeN

teemuk

The only mods to stock values I would consider starting with are: C24 removed and C12 = 10 uF.

You can see their effect in comparison to circuit without mods from the attachment. All tone controls are set to halfway and gain/volume is full. Plots are from the clean channel.

The voltage output may look high but don't worry about it: Rolling the gain control of the clean channel to halfway drops the output voltage below 15V (also seen from the attachment). I assume you don't want to keep your amp at "10" all the time... At output levels "exceeding" the supply's limits you will be just overdriving the power amp or/and the preamp. It is perfectly normal to get some "breakup" from clean channels at higher volume settings.

So... I suggest you start with the initial values and when you have built the thing see if it needs any modifications. At least joecool85 already respects the tone of the "stock" circuit so I see no point in modding it. That, of course, assuming the schematic is correct in the first place, which is why I mentioned the 1uF vs 10uF cap issue. To my eye the 1uF looks too small and like I mentioned, similar amps tend to use higher values there. A higher value also allows the speaker's characteristic resonance to cut through. I believe this is a mistake in the schematic since it seems rather strange to employ current feedback scheme and then damp its most prominent effect. Anyway, unless someone who owns a K-20X opens up his amp and checks out the value of C12 this issue will remain unsolved.

benzer

well
1/ if you want to allow some distortion at high volume levels then u surely cant get rid of the output cap C24 as long as there gonna be some dc component in the distorted output signal (use forrier transforms),and actually it doesnt kill the beauty of the response so i would surely include it, you mentioned that it should be unique somehow, what kind of caps is good to implement as an output cap?

2/ill build the amp with only c12 modified then ill see if it needs another values to be changed.
well i wanna as joecool if any chance he comes here,is the amp introducing distortion at high level clean channel ?specially at high freqs?

3/here i did plot the quarter volume with half bass/treble/midrange boost (output cap is included,c12 is 10uF),above this we start to distort freqs. anyways it looks cool.
the only thing i hate about this response is the low frequencies phase shifts,,its big under 100 hz, hope it doesnt affect much.

maybe after i build this ill start change some values to get my tone ;)

Good OL' BeN

joecool85

As I've already had that amp open numerous times, I don't mind opening it up to check some values for you.  As far as the clean sound goes, I'd have to check it out.  Normally when I'm running it loud it is on the OD channel, or on the clean channel with an OD pedal lol.

I'll check this afternoon for you.
Life is what you make it.
Still rockin' the Dean Markley K-20X
thatraymond.com

benzer

ill be very thankfull man,
,i would realy apreciate it if you can get the right schematic
thanks already
Good OL' BeN