Cristian Ortega del Rio

I’ve started this project like three years ago, collecting information and buying from time to time some electronic supplies, like the lovely LM1875. I wanted to build my first gainclone, this one I’ve made over a year ago, and I’ve never posted here but you can find a thread here about it.

Now, I leave you with some pictures. I sold this unit. The case belonged to a UPS, and was a very small room to use for it.

Parts of my GainClone

Inside the case.

Power supply is made using two EI type xformers 12V @ 5A each and a separated diode bridge with its capacitors 2 x 4700uF per rail.

PSU under PCB

PSU - Component side 2 x 4700uF 25V per rail

It sounds great, and clear. I have a little overheat due to a small heat sink I’m using, but only at max power ratings, and this is really loud. It delivers upto 12W per channel. It’s a very clean sound, I guess if I could add more caps to PSU I could get more clear bass, but It’s ok for this design. I had added an extra 2200uF caps each rail and an extra 470uF as well within POWER AMP PCB, as close as I could from the LMs, take a look inside, and this is the final design and component placement.

Inside my Gainclone

A view under Power Amp PCB - all components on its place

I’m using the same AC filter as the UPS had, this is for power line, and is intended to eliminate noise due to transients under power line, to have an even more clear sound.

I hope you like the design. Here you go more pictures, enjoy them:

And now, the schematic:

Circuit for LM1875

PSU is simple, so I won’t post it here. C1 and C3 are shared for both LM1875 (channels R+L), and their values are those 2200uF + 470uF I refered before.

So, until next update, regards.

I’ve found somewhere on internet this circuit and I thought it would be a nice idea to analyze it. Here it is:

Gain/Attenuation Level Control

This kind of level control circuits allows you to set gain or attenuation levels to adjust any sound level for your audio system. With those values (see schematic) my intention is to be able to set a gain/attenuation from -3.8dB to +3.8dB. And capacitors are chosen to set a BW: 540Hz to 145kHz. This is ideal for placing it after active crossover and before tweeter power amp, and this way you will have a level control for tweeters to adjust any SPL difference between loudspeakers.

Network R8-P1-R11-R10 is used to set amp gain at point R11/R10 (output). Therefore when max gain is set (P1 cursor placed next to R8). Amplifier gain is given by: Av0 = 20 * Log(1+R9/R8) but after that we have a Voltage divider network ((P1+R11) & R10), and as it’s placed series along with the OpAmp, gains/attenuations are added. Therefore the whole system has the following behaviour:

Max Gain: Av0=20 * LOG(1+R9/R8) + 20*LOG((R11+P)/(R10+R11+P))

Max Att: Av1 = 20 * LOG(1+R9/(R8+P)) + 20 * LOG(R11/(R10+R11))

Reordering, and using Log property:

Av0= 20 * LOG((1+R9/R8) * ((R11+P)/(R10+R11+P)))

Av1= 20 * LOG((1+R9/(R8+P)) * R11/(R10+R11))

Why are we using a Voltage divider on the opamp output?

Well, due to its non inverting configuration the minimum gain we can get is 0dB (Av=1), therefore if we want a negative dB gain (attenuation) we have to use that Voltage divider for Av<1, and that’s the explanation I think fits better to that question.

Well, it has been a long time since last update, I will keep posting more often from now on.

Regards.

I was told, and It’s a common belief that woofer’s diameter is directly proportional to wavelength, or inversely proportional to frequency. And that explains why woofer speaker are lot bigger than tweeters.

Well, that’s not true there’s nothing related between both parameters (except for system’s transient response, that maybe the resultant srping-mass subsystem cone-suspention is in part affected in different way depending on cone diameter, I don’t think so either). Looking for an answer on internet I’ve found a great article here, well explained you can find answers there.

But I have my own opinion based upon info found mostly in that article, but also what I do know about SPL, power amp and speaker diameter relations.

So, we can state that: wavelength has nothing to do with woofer’s (speaker) diameter. It’s all matter about SPL, and nothing more than that. Let’s take an example of two woofers: one 8” diameter, the other 10” diameter. If both have the same Xmax (maximum excursion), it is obvious that 10” woofer will “move” more air volume than 8”, how much? Well as this volume of air is related to diameter squared, a 10” woofer will move about 50% more air than 8” woofer.

Therefore under same conditions, SPL of a 10” is bigger than 8” (attached to a same sealed or vented box, at least 3dB different). As you will need more (or same) power for bass than mid+high, it would seems (actually it will do) that a 10” “sounds more” than 8” woofer. And you can hear better lower frequencies. Therefore someone could believe that a bigger woofer could go lower than a small one, because those low frequencies sound louder.

Whenever you want to buy a nice speaker you will take into account several parameters, not only RMS and Peak power, BW and Fs, but also max SPL, and Xmax at least. So, could be better to install a bigger woofer to a loudspeaker system? Yes. How big? As much as your pocket and requirements can afford. What is a nice diameter as an advice? A 15” in my opinion will do a great deal! My pocket right now can afford a nice 10” LOL.

Remember only when we talk about Radiofrequency we can take into account that ratio between wavelenght and driver size, in this case the size of the Antenna. No matter which configuration, dipole, yagui, parabolics, etc. The higher the frequency, the lower the lenght of the antenna.

So, until next update!

Due to a lack of tools problem, I have to stop pluto design project . However, I will work right away into a new project, a brand new baffle. I will use some ideas from pluto electronics design. Soon will have more news.
About my pluto, I leave you with some pictures of it, at least as much as I could come up with.

All this effort is not good enough to finish a pluto, but is worth the try to learn how to build one. Because i needed to cut more wood, to build a pipe of wood. My idea, was to cut several rings of wood and then glue them all together in order to create this pipe and adapt the PVC pipe (110mm Ø) with loudspeaker (150mm Ø), but that lack of tools  stops me for now. I don’t have a propper tool for it.

This is what I’ve got so far. And that’s all folks! For now this is what I have, around 35% finished.

So, until next update! .·.

From the analysis made on last two posts we have the following circuit:

Input buffer circuit

This circuit is taken from Pluto Electronics page, and we have made a previous analysis, therefore now we’re going to set its BOM (Bill Of Materials), and we are going to simulate it and see how it works in frequency domain, with a slightly difference from linkwitz’s design, to fit my circuit. Also I have different speakers, therefore the whole system will be different. Well, this would be my B.O.M.

Resistors

R1= 1K
R2= 100
R3= 1.96K
R4= 3.48K
R5= 1K
R6= 33.2K

Capacitors

C1= 2.2 uF
C2= 220 pF
C3= 220 nF
C4= 560pF

Operational Amplifier

Your choice: NE5532, OPA627, even TL082

With these values, we can get a -3dB freq as follows LP: 37Hz, HP: 208kHz for active stage, and for low input signal, that passive stage would be LP: 22Hz and HP: 285kHz aprox. I have chosen those values for components because I already have most of them.

You may wonder: why are they realitively small values? Well I think you should ask yourself why give noise a chance? Don’t you think?

This below is a Bode Simulation made by Multisim software.

Bode Analysis for this buffer design - Low Level Input Simulation

In this case I have used a NE5532 opamp. All resistors 1% metal film. And for High level Input, this below there is its Bode simulation.

Bode Simulation for High Level Input.

As you can see they are almost the same shape, which is good for our purposes.

Well, I will try to do some experiments and analysis, tests etc, using this circuit in order to find out more about it. Please feel free to share your experiences about this or any other configuration. I will do some research about input buffers and all that. I have many ideas to share.

I’m still working on PLUTO design, my god! I hope to finish some day.

Take care!

This is the second part of Input Buffer post.

And we continue with the analisys, the part of circuit for a High level input is as follows.

Fig 1 - High Level Input buffer

As you can see at first stage we have a voltage divider, in order to adapt its input signal to a proper value, recommended values are in a ratio 10:1, to get a signal V’i at 9% of original input. As we did on its related post, we will analyze this circuit under two circumstances, first a low freq value, and then a high freq value. Take into account now that this configuration is an inverting op-amp, so its gain formula is:

Eq 1.01 - Gain and Transfer Function eqs.

First case, low frequency, capacitor C2 has no influence on R4, therefore Z2(s) is equal to R4. This is due to cap’s low value, the order of pF so its impedance trends to infinit.

Eq 1.02 - Transfer function HP Filter section and C1 formula.

Now in the case of a higer value for frequency, impedance of C1 (XC1) trends to zero, then only R3 is important in this case, therefore C1 won’t affect our circuit and can be dismissed from our analysis.

Eq 1.03 - C2 Calculation.

And these are the whole analysis of this first stage: input buffer section, for high and low input signals. Next time I will use all these equations to build a sample circuit of an input buffer, a circuit that will be used on my PLUTO electronics.

Bibliography:

• Enciclopedia de la Electrónica, Cap. 23, pag. 652 – C. Belove.

I’m working on a nice project, a Linkwitz’s idea. I’m working on his PLUTO speaker system which is great and easy to DIY, I’m reading the whole site and I have found a very interesting proposal along with pluto’s electronics.

I have been analyzing all parts and I thought that would be really neat to write something about each one of them. I’m not going to use LM3886 ICs, instead I will use LM1875, therefore my project is smaller and also a slightly different than his. I should start from the PSU, instead of that I will start writing about its input buffer, filters (crossovers) and power amp section, and after that I will focus on PSU. You may wonder why LM1875? Well simple, because I already have those, and because i love them. And most important because I live in an apartment, so I do this for my neighbors’ sake.

Well, let’s talk about buffering, one thing i find a pretty neat trick is the abilty for Hi and Lo Level Inputs, and I will focus on these two options, because I will use them , below there’s a standard input buffer with Hi and Lo inputs.

Let’s examinate the Low Level Input first. We can simplify this circuit above, then, by the following schematic.

On the very first stage we found a High Pass Filter (F HP1), made by C3 with R5 and R6, assuming that C3 >> C4, therefore C4 with R6 and R5 make a Low Pass Filter (F LP1). For fh1, fh2, fl1 and fl2 calculations (cutoff frequencies at -3dB) we do the following:

Given: R5, R6 and fhp1 (high pass frequency cutoff)

if R6 >> R5, then we can assume:

Given a flp1 (low pass freq cutoff)

Then we calculate buffer gain, Av, and after that we calculate its cutoff frequencies. We have to take into account that dB gain for power signals is given by: A[dB] = 10 * LOG(Ap), therefore in order to find out this fc at -3dB, we need to assume that Av[-3dB] => Av/(2)^0.5.  So, Av[dB] = 20 * LOG(Av) , and Av[-3dB] = 20 * LOG(Av * 0.707)

From Fig 2 we get this forumla:

Now we have another stage, OA stage, with its gain as shown above, but it also has pasive components,  same as its input stage as a low pass and a high pass filters. And that makes it to behave as a bandpass filter. So we can evaluate and calculate all variables involved, but first there are some considerations before picking up components values. We should consider two options to understand how this circuit will work on frecuency domain. First f>0Hz (low freq) and after that f>>0Hz (high freq). But first let’s consider the following schematic, for our calculus purposes.

From the schematic above, we can get the following:

First case when f is almost0Hz, low frequency signal input, that means a few decades of Hz and we consider C1 connected therefore Xc1 > 0ohms. And due to small value of C2, its impedance is very high and has no effect upon R4 , therefore:

Eq 1.06 - Calculus of Rx and therefore Xc1 and C1

Same criteria goes for C2 calculation. Now we’re talking about higher freq, fc>>0Hz, therefore we must consider that Xc1=0ohms, but Xc2 is not infinite anymore, and has influence on R4, then the amplifier’s gain changes as well. In order to estimate its cutoff frequency, we must find (again) its transfer function (G(s)) and its poles, then analyze them in frequency domain aspect.

Eq 1.08 - Calculus for C2

And that’s it, for now. This is how we can set all parameters for an input buffer, of course this is only one part, the low level input on my next post i will talk about the high level input, that means that I will only talk about how it behaves due to all these parameters found here. There will be no search for Caps or Resistors values at all. But maybe we can come into a compromise with all calculated and estimated values here and there.

Finally i leave you with a spreadsheet that makes all these calculations easier, fields in yellow background are values we must enter.

Bibliography:

• Sistemas de Control Automático – Cap 9: 9-1-2 Especificaciones en el dominio de la frecuencia, pag. 543 – Benjamin Kuo.
• http://en.wikipedia.org/wiki/Bandwidth_%28signal_processing%29
• http://en.wikipedia.org/wiki/Gain
• http://en.wikipedia.org/wiki/Low-pass_filter
• and of course, Linkwitz’s website.

Thanks, until next update.

Welcome to Cristian Ortetga del Rio’s Site.