This is a good read and from Eric himself.....the second last sentence suport my tests as well.
""A PI relies on a strong current pulse into the coil to obtain maximum depth. The more current into the coil the greater the potential of obtaining more depth. Next a very low noise front end of the detector is required so any return signal can be noticed. This takes a lot of engineering time and energy to maximize this feature.
So, as a general rule, high power PI's go deeper and those that have the greatest development in reducing front end noise also have a tendency to get the greatest depth.
Typical high power PI's include the ML PI series, Eric Foster's Goldscan series, and the Whites TDI. There are others, but these are the most often referenced. These detectors use heavy batteries so they can provide the greater current needed to maximize depth.
Lower powered PI's such as the Beachscan, Goldquest, Headhunter (HH), Whites dual field (DF), C Scope CS 6 and some of their other PI series, as well as other brands using typical AA batteries are usually a low power or lower powered than the above mentioned high powered PI's. Normally, these detectors have a current limiting resistor in series with the coil to limit the current, which in turn extends battery life.
Many of the low powered designs used today or even those of the past are detectors originally designed by Eric Foster. The evolution of some follow closely to Beachscan. The Headhunter is one of them that I believe was built between the Beachscan and the Goldquest. All of these are low powered PI's and most, but not all use, 8 or so AA batteries.
Now, the HH uses two 9v batteries. This is more voltage than that produced by 8 AA's but the two 9V cannot and do not produce nearly as much current into the coil as the AA battery pack and it is the current into the coil and not the voltage that is what is needed to produce the stronger pulse. The two 9v batteries was chose to provide the necessary voltage for much of the control circuitry and not to provide more coil current and do so at a lighter weight. In fact, I suspect a larger value resistor is installed in the coil circuitry to reduce the coil current than found on the Beachscan series. Also, keep in mind that is is a larger pulse current is what is needed to increase depth.
Keep in mind the Minelab PI series uses a heavy duty 6 volt battery that also uses more pulse current than other PI's that I know of even though the voltage is less than most others including the GS 5 and TDI. Also keep in mind that the ML PI also goes the deepest and does so because of both the increased current into the coil and the advanced engineering in reducing noise generated in the front end of the electronics. This advanced depth capability is also reflected in the price of the detector.
So, in a nutshell, the operating voltage by itself has little to do with just how deep a PI will go.
Now, discussing the low powered PI's such as the HH, the Beachscan, the Goldquest, and others, the low power label has little to do with the depth potential. Yes, they do not put the largest pulse into the coil, but the increased repetition rate helps in the depth potential. No, the low powered PI's won't go as deep as the high powered PI's but the depth difference isn't as great as one might think.
One other factor that applies is the fact that the technique presently used to ground balance a detector such that one can ignore or cancel the ground mineral signals is done by adding more samples taken later in time and then subtracting the later signals that have been amplified enough such that the signal from the ground in the later sample equal the main sample and the later is subtracted from the main sample. The result is the ground signal appears to go away. What also goes away is part of the signal from many targets. Thus, ground balancing detectors will not have the depth on many of the same objects as a straight PI of similar design not using or having the ground balance feature. This is why on the TDI, turning off the ground balance (GB) in areas where it can be done results in more depth on most objects. This is also why many TDI owners complain that they can see an obvious depth loss on some objects when testing with the GB on versus GB off.
I mentioned the low powered PI's do quite well against other PI's and part of the reason is the fact they do not have the ground balance system. So, they do not suffer the depth loss caused by using a GB design.
The bottom line is there are a lot of factors involved as to what affects the depth potential of a PI detector. Also keep in mind, there are anomalies not related to the detector that can influence the potential depth one might be able to detect an object. Normally, this occurs on a one time basis and is not the norm. Normal repeatable depth capability is generally much less. However, we as humans demand our bragging rights and take advantage of them. So, it is not unusual to read of abnormal depth capabilities being posted. This is far more common on some of the gold prospecting forums where one can easily picture a fisherman's ruler being used.
Now, with all the above mentioned info I will also mention that something as simple as EMI (external electrical noise) plays a large part on what one might be able to detect an object, so testing should be done often and at different locations to really get an honest evaluation. While building coils I ran into some strange things that stumped me at first and one of those things was the noise factor. I found while tweaking coils I could get depths ranging from about 11" to a maximum of 17" air test on a nickel. Now, that is a huge variation that caused me some problems because it wasn't consistent. So, coil testing became a challenge. The final result was there was very little depth difference between two similar size and shape coils once all the other factors were equal. Oh yeah, the realistic nickel depth of detection on my low powered PI's was about 11" to 12" and that was on a good day.
I hope this clears up some of the errors in thinking many people may have as well as why one can read such dramatic differences in postings about depth of detection."